YOLOV3_colab.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "YOLOV3.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "5RIzKhSfNV8W",
        "MQxHNRCvNY0K",
        "NxLPUSN-NdzB",
        "uP_FzHJuNjW7",
        "w2QS7yUzNmUm",
        "ybcjjd5PpknP"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "VtcyVYYKLU4r",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 136
        },
        "outputId": "b21095d4-55cc-42b7-bb5e-938ce5d1afbe"
      },
      "source": [
        "!git clone https://github.com/pythonlessons/TensorFlow-2.x-YOLOv3.git"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'TensorFlow-2.x-YOLOv3'...\n",
            "remote: Enumerating objects: 120, done.\u001b[K\n",
            "remote: Counting objects:   0% (1/120)\u001b[K\rremote: Counting objects:   1% (2/120)\u001b[K\rremote: Counting objects:   2% (3/120)\u001b[K\rremote: Counting objects:   3% (4/120)\u001b[K\rremote: Counting objects:   4% (5/120)\u001b[K\rremote: Counting objects:   5% (6/120)\u001b[K\rremote: Counting objects:   6% (8/120)\u001b[K\rremote: Counting objects:   7% (9/120)\u001b[K\rremote: Counting objects:   8% (10/120)\u001b[K\rremote: Counting objects:   9% (11/120)\u001b[K\rremote: Counting objects:  10% (12/120)\u001b[K\rremote: Counting objects:  11% (14/120)\u001b[K\rremote: Counting objects:  12% (15/120)\u001b[K\rremote: Counting objects:  13% (16/120)\u001b[K\rremote: Counting objects:  14% (17/120)\u001b[K\rremote: Counting objects:  15% (18/120)\u001b[K\rremote: Counting objects:  16% (20/120)\u001b[K\rremote: Counting objects:  17% (21/120)\u001b[K\rremote: Counting objects:  18% (22/120)\u001b[K\rremote: Counting objects:  19% (23/120)\u001b[K\rremote: Counting objects:  20% (24/120)\u001b[K\rremote: Counting objects:  21% (26/120)\u001b[K\rremote: Counting objects:  22% (27/120)\u001b[K\rremote: Counting objects:  23% (28/120)\u001b[K\rremote: Counting objects:  24% (29/120)\u001b[K\rremote: Counting objects:  25% (30/120)\u001b[K\rremote: Counting objects:  26% (32/120)\u001b[K\rremote: Counting objects:  27% (33/120)\u001b[K\rremote: Counting objects:  28% (34/120)\u001b[K\rremote: Counting objects:  29% (35/120)\u001b[K\rremote: Counting objects:  30% (36/120)\u001b[K\rremote: Counting objects:  31% (38/120)\u001b[K\rremote: Counting objects:  32% (39/120)\u001b[K\rremote: Counting objects:  33% (40/120)\u001b[K\rremote: Counting objects:  34% (41/120)\u001b[K\rremote: Counting objects:  35% (42/120)\u001b[K\rremote: Counting objects:  36% (44/120)\u001b[K\rremote: Counting objects:  37% (45/120)\u001b[K\rremote: Counting objects:  38% (46/120)\u001b[K\rremote: Counting objects:  39% (47/120)\u001b[K\rremote: Counting objects:  40% (48/120)\u001b[K\rremote: Counting objects:  41% (50/120)\u001b[K\rremote: Counting objects:  42% (51/120)\u001b[K\rremote: Counting objects:  43% (52/120)\u001b[K\rremote: Counting objects:  44% (53/120)\u001b[K\rremote: Counting objects:  45% (54/120)\u001b[K\rremote: Counting objects:  46% (56/120)\u001b[K\rremote: Counting objects:  47% (57/120)\u001b[K\rremote: Counting objects:  48% (58/120)\u001b[K\rremote: Counting objects:  49% (59/120)\u001b[K\rremote: Counting objects:  50% (60/120)\u001b[K\rremote: Counting objects:  51% (62/120)\u001b[K\rremote: Counting objects:  52% (63/120)\u001b[K\rremote: Counting objects:  53% (64/120)\u001b[K\rremote: Counting objects:  54% (65/120)\u001b[K\rremote: Counting objects:  55% (66/120)\u001b[K\rremote: Counting objects:  56% (68/120)\u001b[K\rremote: Counting objects:  57% (69/120)\u001b[K\rremote: Counting objects:  58% (70/120)\u001b[K\rremote: Counting objects:  59% (71/120)\u001b[K\rremote: Counting objects:  60% (72/120)\u001b[K\rremote: Counting objects:  61% (74/120)\u001b[K\rremote: Counting objects:  62% (75/120)\u001b[K\rremote: Counting objects:  63% (76/120)\u001b[K\rremote: Counting objects:  64% (77/120)\u001b[K\rremote: Counting objects:  65% (78/120)\u001b[K\rremote: Counting objects:  66% (80/120)\u001b[K\rremote: Counting objects:  67% (81/120)\u001b[K\rremote: Counting objects:  68% (82/120)\u001b[K\rremote: Counting objects:  69% (83/120)\u001b[K\rremote: Counting objects:  70% (84/120)\u001b[K\rremote: Counting objects:  71% (86/120)\u001b[K\rremote: Counting objects:  72% (87/120)\u001b[K\rremote: Counting objects:  73% (88/120)\u001b[K\rremote: Counting objects:  74% (89/120)\u001b[K\rremote: Counting objects:  75% (90/120)\u001b[K\rremote: Counting objects:  76% (92/120)\u001b[K\rremote: Counting objects:  77% (93/120)\u001b[K\rremote: Counting objects:  78% (94/120)\u001b[K\rremote: Counting objects:  79% (95/120)\u001b[K\rremote: Counting objects:  80% (96/120)\u001b[K\rremote: Counting objects:  81% (98/120)\u001b[K\rremote: Counting objects:  82% (99/120)\u001b[K\rremote: Counting objects:  83% (100/120)\u001b[K\rremote: Counting objects:  84% (101/120)\u001b[K\rremote: Counting objects:  85% (102/120)\u001b[K\rremote: Counting objects:  86% (104/120)\u001b[K\rremote: Counting objects:  87% (105/120)\u001b[K\rremote: Counting objects:  88% (106/120)\u001b[K\rremote: Counting objects:  89% (107/120)\u001b[K\rremote: Counting objects:  90% (108/120)\u001b[K\rremote: Counting objects:  91% (110/120)\u001b[K\rremote: Counting objects:  92% (111/120)\u001b[K\rremote: Counting objects:  93% (112/120)\u001b[K\rremote: Counting objects:  94% (113/120)\u001b[K\rremote: Counting objects:  95% (114/120)\u001b[K\rremote: Counting objects:  96% (116/120)\u001b[K\rremote: Counting objects:  97% (117/120)\u001b[K\rremote: Counting objects:  98% (118/120)\u001b[K\rremote: Counting objects:  99% (119/120)\u001b[K\rremote: Counting objects: 100% (120/120)\u001b[K\rremote: Counting objects: 100% (120/120), done.\u001b[K\n",
            "remote: Compressing objects: 100% (119/119), done.\u001b[K\n",
            "remote: Total 602 (delta 68), reused 0 (delta 0), pack-reused 482\u001b[K\n",
            "Receiving objects: 100% (602/602), 70.75 MiB | 30.59 MiB/s, done.\n",
            "Resolving deltas: 100% (337/337), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B99c-h76LktG",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "c8307d83-6a2b-48ae-f4d2-dc0060bcc275"
      },
      "source": [
        "%cd TensorFlow-2.x-YOLOv3/"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/TensorFlow-2.x-YOLOv3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MZ3lDT2WLpF2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "721605aa-ae02-474d-aed7-b2e186d4d06f"
      },
      "source": [
        "!python mnist/make_data.py"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Extracting all train files now...\n",
            "Done!\n",
            "Extracting all test files now...\n",
            "Done!\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000001.jpg 179,27,193,41,1 242,272,284,314,0 322,187,406,271,2 120,195,232,307,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000002.jpg 385,295,399,309,0 14,241,36,263,6 327,4,349,26,9 328,59,370,101,6 179,243,221,285,1 172,188,214,230,2 175,45,217,87,1 245,103,357,215,1 40,265,152,377,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000003.jpg 69,337,91,359,9 215,281,229,295,4 162,296,218,352,2 324,49,352,77,3 49,263,77,291,5 349,173,391,215,3 21,66,133,178,4 284,255,396,367,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000004.jpg 76,372,98,394,6 134,296,176,338,1 377,269,405,297,7 249,139,277,167,8 348,176,404,232,2 371,31,399,59,2 184,207,240,263,1 255,291,367,403,2 44,124,156,236,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000005.jpg 211,351,233,373,9 308,18,330,40,0 324,369,338,383,3 228,49,312,133,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000006.jpg 256,107,312,163,9 97,91,139,133,5 239,361,267,389,4 27,111,83,167,3 46,204,74,232,9 232,195,344,307,5 60,291,172,403,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000007.jpg 76,117,90,131,4 212,222,226,236,8 366,58,394,86,3 253,341,281,369,6 135,24,219,108,2 22,65,134,177,9 266,75,350,159,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000008.jpg 353,282,375,304,4 217,351,231,365,2 201,284,229,312,5 94,58,150,114,3 98,225,140,267,4 250,37,278,65,0 170,71,226,127,2 227,79,283,135,9 2,254,114,366,6 184,133,268,217,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000009.jpg 333,386,355,408,4 311,96,325,110,2 166,347,188,369,9 178,149,220,191,4 146,83,202,139,3 19,41,47,69,0 224,110,280,166,1 302,239,414,351,2 135,205,219,289,1 11,226,123,338,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000010.jpg 338,176,394,232,3 285,81,341,137,2 201,330,285,414,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000011.jpg 55,334,97,376,9 185,5,297,117,8 40,153,124,237,0 66,1,178,113,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000012.jpg 149,106,163,120,9 386,131,408,153,9 299,379,321,401,1 322,286,378,342,2 310,92,366,148,8 278,178,334,234,9 139,217,195,273,9 21,277,77,333,9 90,171,118,199,6 89,274,173,358,1 34,28,146,140,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000013.jpg 169,78,191,100,2 335,360,349,374,3 380,98,408,126,3 111,130,139,158,3 12,52,54,94,3 114,180,198,264,5 99,311,183,395,8 240,127,324,211,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000014.jpg 269,392,283,406,5 131,368,173,410,0 8,271,64,327,3 280,194,392,306,6 124,102,236,214,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000015.jpg 326,245,382,301,8 254,88,282,116,7 125,307,167,349,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000016.jpg 25,171,47,193,1 183,177,197,191,5 26,95,40,109,6 119,256,147,284,9 124,331,152,359,3 313,263,341,291,1 212,318,268,374,9 202,66,314,178,7 93,24,205,136,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000017.jpg 356,172,370,186,9 217,207,231,221,9 341,201,397,257,0 245,163,301,219,5 283,276,325,318,3 153,324,209,380,1 218,240,260,282,1 263,37,291,65,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000018.jpg 215,61,229,75,9 24,223,52,251,6 160,27,202,69,8 316,206,344,234,6 152,270,180,298,3 190,102,246,158,6 37,266,79,308,3 272,237,384,349,7 96,294,208,406,1 0,96,112,208,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000019.jpg 359,313,401,355,6 213,91,241,119,8 144,215,256,327,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000020.jpg 130,207,152,229,2 394,342,408,356,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000021.jpg 308,372,330,394,2 325,35,347,57,6 280,163,322,205,9 303,279,359,335,3 97,28,209,140,7 206,254,290,338,8 7,243,119,355,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000022.jpg 80,51,94,65,3 59,21,87,49,0 288,282,344,338,7 128,191,212,275,2 303,164,415,276,0 167,61,279,173,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000023.jpg 276,201,318,243,0 46,180,102,236,5 181,238,209,266,2 323,294,351,322,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000024.jpg 27,302,49,324,4 128,204,170,246,1 209,139,321,251,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000025.jpg 203,65,225,87,3 108,164,130,186,8 276,401,290,415,5 35,156,91,212,1 164,12,192,40,0 26,100,82,156,0 250,42,292,84,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000026.jpg 36,145,50,159,2 320,292,348,320,3 344,218,400,274,2 101,148,213,260,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000027.jpg 258,231,300,273,8 104,2,146,44,5 292,133,348,189,8 174,140,216,182,2 191,53,247,109,1 47,204,131,288,6 297,215,381,299,5 146,260,258,372,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000028.jpg 49,207,63,221,6 106,337,128,359,1 37,370,65,398,0 171,241,213,283,5 225,43,267,85,4 294,198,350,254,1 123,232,165,274,7 21,8,105,92,1 75,125,187,237,6 215,290,327,402,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000029.jpg 296,393,310,407,4 167,334,181,348,9 40,108,68,136,3 44,160,86,202,7 73,36,129,92,5 211,182,239,210,4 79,308,135,364,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000030.jpg 353,252,381,280,4 46,192,88,234,1 295,34,323,62,8 277,89,389,201,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000031.jpg 63,276,85,298,3 302,110,316,124,0 114,11,128,25,2 195,254,237,296,5 368,308,410,350,1 143,23,185,65,9 264,255,292,283,6 194,34,222,62,0 102,100,144,142,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000032.jpg 79,300,101,322,0 214,124,236,146,9 152,216,194,258,1 197,254,239,296,0 271,30,299,58,6 304,106,388,190,4 18,64,130,176,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000033.jpg 280,298,308,326,4 257,188,313,244,9 158,305,186,333,6 284,112,326,154,5 64,49,92,77,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000034.jpg 197,48,219,70,4 17,253,39,275,6 35,6,63,34,5 11,71,67,127,1 41,130,97,186,6 120,41,176,97,1 303,194,331,222,2 199,297,311,409,0 84,256,196,368,5 248,74,360,186,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000035.jpg 277,129,291,143,0 47,294,75,322,9 222,260,278,316,6 268,347,310,389,3 170,129,226,185,2 108,282,150,324,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000036.jpg 285,185,313,213,5 293,46,349,102,7 291,100,319,128,3 136,209,164,237,5 0,361,28,389,2 203,91,287,175,7 195,237,307,349,0 185,5,269,89,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000037.jpg 207,189,229,211,3 87,26,109,48,5 238,293,266,321,7 41,239,69,267,1 334,138,390,194,6 96,369,124,397,1 322,42,378,98,2 228,126,312,210,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000038.jpg 289,224,311,246,4 344,343,400,399,9 118,130,146,158,2 314,183,370,239,8 295,356,337,398,2 102,266,158,322,4 169,70,281,182,6 183,235,267,319,7 292,75,404,187,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000039.jpg 194,94,250,150,4 129,322,185,378,5 139,214,195,270,3 264,38,320,94,8 268,188,380,300,8 47,63,159,175,6 30,191,142,303,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000040.jpg 382,207,404,229,7 379,276,393,290,3 97,151,153,207,6 155,202,211,258,7 251,267,279,295,3 271,84,327,140,5 32,126,60,154,2 67,47,151,131,3 169,115,253,199,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000041.jpg 390,215,404,229,6 167,210,181,224,8 392,381,414,403,0 201,278,285,362,4 104,119,216,231,3 107,22,191,106,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000042.jpg 246,66,268,88,9 88,117,102,131,7 107,148,121,162,6 108,223,150,265,4 251,233,279,261,6 120,162,176,218,4 272,322,300,350,0 301,298,329,326,3 305,122,347,164,5 168,253,252,337,1 323,200,407,284,7 103,15,215,127,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000043.jpg 51,166,65,180,8 343,353,399,409,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000044.jpg 173,274,229,330,9 211,79,323,191,6 89,236,173,320,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000045.jpg 174,285,230,341,0 83,214,125,256,0 171,106,283,218,4 301,84,413,196,4 227,8,311,92,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000046.jpg 255,237,277,259,6 202,222,216,236,7 198,314,220,336,1 281,183,309,211,2 361,144,389,172,7 338,330,380,372,7 332,263,388,319,5 275,108,303,136,1 9,110,93,194,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000047.jpg 156,248,178,270,7 39,148,67,176,2 300,237,328,265,5 180,322,222,364,7 232,207,288,263,0 354,191,382,219,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000048.jpg 158,121,180,143,8 282,313,304,335,8 115,321,143,349,3 84,372,126,414,1 234,223,262,251,8 320,158,362,200,6 364,215,406,257,8 24,150,52,178,1 253,77,337,161,8 147,273,259,385,2 61,66,145,150,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000049.jpg 184,355,206,377,9 325,2,339,16,6 334,28,390,84,8 307,234,335,262,9 43,256,155,368,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000050.jpg 197,123,281,207,2 163,5,275,117,4 89,185,201,297,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000051.jpg 304,205,318,219,9 171,193,185,207,8 43,150,65,172,1 142,365,170,393,3 298,110,340,152,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000052.jpg 29,5,43,19,6 209,225,237,253,1 60,9,116,65,6 65,222,93,250,7 297,225,409,337,3 88,318,172,402,5 22,122,106,206,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000053.jpg 283,21,305,43,8 208,281,264,337,4 202,147,244,189,2 278,108,334,164,6 9,220,121,332,1 294,293,406,405,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000054.jpg 123,283,145,305,7 23,213,37,227,7 210,364,224,378,7 213,207,297,291,8 44,90,156,202,1 320,275,404,359,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000055.jpg 225,68,247,90,4 327,169,349,191,6 145,283,173,311,9 302,225,358,281,3 115,177,171,233,1 259,338,287,366,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000056.jpg 166,139,180,153,2 23,17,65,59,0 0,153,56,209,8 24,270,80,326,3 309,39,351,81,7 344,269,372,297,8 321,304,363,346,4 47,81,131,165,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000057.jpg 343,380,357,394,3 68,130,180,242,3 71,308,155,392,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000058.jpg 22,58,50,86,4 221,267,263,309,0 273,250,329,306,4 127,103,155,131,0 264,158,348,242,6 38,177,122,261,2 284,59,368,143,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000059.jpg 69,74,91,96,3 236,69,264,97,9 73,168,101,196,4 30,322,114,406,8 259,103,371,215,8 321,6,405,90,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000060.jpg 34,172,48,186,1 105,125,127,147,3 13,209,35,231,6 262,87,290,115,8 315,289,357,331,5 109,33,151,75,5 76,180,118,222,2 271,384,299,412,0 171,129,283,241,8 104,302,188,386,7 22,41,106,125,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000061.jpg 260,323,274,337,3 52,316,74,338,0 278,326,320,368,0 177,335,233,391,9 332,161,388,217,3 49,85,105,141,2 207,77,319,189,6 276,212,388,324,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000062.jpg 308,349,330,371,7 316,120,330,134,1 103,151,215,263,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000063.jpg 399,51,413,65,9 17,382,45,410,6 74,229,116,271,6 99,54,211,166,8 303,18,415,130,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000064.jpg 366,42,388,64,2 365,362,379,376,9 63,206,105,248,3 270,349,298,377,7 333,264,375,306,0 264,271,306,313,6 338,331,366,359,1 147,268,231,352,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000065.jpg 232,69,344,181,8 40,234,124,318,2 295,204,407,316,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000066.jpg 392,263,414,285,0 369,369,391,391,8 24,259,80,315,7 183,156,295,268,8 293,57,405,169,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000067.jpg 87,298,101,312,4 67,155,89,177,3 119,5,161,47,6 101,124,157,180,6 283,0,339,56,7 263,241,291,269,2 321,127,377,183,0 231,325,315,409,1 50,212,162,324,2 169,155,253,239,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000068.jpg 174,237,188,251,9 169,301,211,343,1 78,373,120,415,1 279,160,335,216,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000069.jpg 224,12,252,40,0 211,373,253,415,4 31,114,59,142,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000070.jpg 245,309,267,331,2 253,153,275,175,3 170,266,226,322,2 357,24,413,80,7 364,337,406,379,9 153,4,237,88,5 217,182,329,294,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000071.jpg 19,372,41,394,4 119,82,133,96,6 27,1,83,57,8 99,357,141,399,8 133,162,161,190,6 258,213,314,269,4 20,59,48,87,9 200,17,242,59,1 294,86,406,198,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000072.jpg 385,130,399,144,8 276,67,304,95,9 239,225,281,267,0 274,145,316,187,5 61,93,89,121,5 358,283,400,325,1 315,378,343,406,8 125,164,209,248,9 314,56,398,140,5 133,272,245,384,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000073.jpg 16,197,30,211,7 64,95,78,109,7 146,197,168,219,4 301,336,357,392,0 24,152,80,208,1 317,85,359,127,1 172,22,200,50,1 120,65,232,177,4 199,227,283,311,6 53,259,137,343,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000074.jpg 101,374,123,396,4 398,331,412,345,6 141,108,253,220,4 308,278,392,362,9 329,54,413,138,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000075.jpg 162,308,184,330,1 137,195,165,223,3 256,153,312,209,9 100,116,142,158,4 290,36,402,148,9 2,149,114,261,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000076.jpg 37,81,93,137,8 180,353,236,409,2 150,319,192,361,6 74,133,102,161,2 309,226,365,282,6 233,289,345,401,6 140,72,224,156,5 74,229,158,313,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000077.jpg 315,155,337,177,0 389,328,403,342,6 31,267,115,351,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000078.jpg 375,204,397,226,9 50,18,106,74,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000079.jpg 194,212,216,234,0 366,211,380,225,9 58,341,72,355,5 137,272,221,356,4 148,47,232,131,0 227,275,339,387,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000080.jpg 124,240,180,296,4 259,377,287,405,3 255,135,283,163,1 157,8,185,36,3 190,290,218,318,7 328,200,370,242,5 30,78,142,190,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000081.jpg 273,49,295,71,4 88,258,116,286,7 312,22,368,78,0 349,280,391,322,6 368,185,396,213,9 77,12,133,68,9 211,142,267,198,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000082.jpg 81,321,95,335,0 108,122,130,144,3 329,313,343,327,8 317,90,359,132,9 152,270,236,354,2 258,187,370,299,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000083.jpg 291,78,313,100,0 386,44,408,66,3 43,23,57,37,0 3,317,59,373,9 301,246,385,330,4 128,105,212,189,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000084.jpg 114,227,136,249,1 6,124,62,180,9 303,65,359,121,9 78,52,162,136,3 81,271,165,355,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000085.jpg 215,212,229,226,2 355,324,411,380,0 280,156,322,198,5 70,194,154,278,5 331,93,415,177,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000086.jpg 187,144,243,200,9 141,60,225,144,5 302,204,386,288,1 157,217,269,329,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000087.jpg 205,33,219,47,3 147,247,169,269,8 147,281,161,295,8 284,255,312,283,6 320,95,376,151,6 34,223,62,251,0 247,141,289,183,5 311,286,367,342,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000088.jpg 46,121,68,143,9 198,64,212,78,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000089.jpg 364,369,386,391,3 344,14,366,36,0 343,263,385,305,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000090.jpg 97,227,119,249,3 384,279,398,293,6 197,275,225,303,2 254,263,296,305,2 345,139,387,181,1 124,40,152,68,4 243,4,285,46,7 122,77,234,189,4 4,49,116,161,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000091.jpg 48,153,70,175,3 199,32,241,74,9 127,90,169,132,0 176,308,204,336,0 248,180,290,222,4 23,290,51,318,3 93,193,149,249,2 242,263,354,375,4 63,269,175,381,6 293,14,377,98,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000092.jpg 169,91,211,133,3 176,243,232,299,4 381,9,409,37,0 265,263,307,305,5 322,272,406,356,7 31,267,115,351,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000093.jpg 385,263,407,285,2 321,343,343,365,1 64,393,78,407,9 384,305,412,333,4 270,230,298,258,7 164,255,206,297,5 316,189,358,231,5 91,226,147,282,6 31,314,59,342,5 16,22,128,134,1 172,116,256,200,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000094.jpg 294,83,322,111,5 255,195,311,251,7 211,259,323,371,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000095.jpg 109,255,131,277,8 324,9,366,51,6 168,107,252,191,7 21,29,133,141,3 49,144,161,256,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000096.jpg 230,158,244,172,0 161,270,175,284,1 119,149,175,205,8 101,310,157,366,2 177,232,219,274,7 213,198,241,226,7 147,39,259,151,1 210,267,322,379,0 303,41,415,153,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000097.jpg 89,54,103,68,5 46,312,102,368,6 338,219,366,247,6 299,112,355,168,9 193,211,221,239,1 99,188,141,230,2 333,327,361,355,4 212,69,296,153,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000098.jpg 215,198,237,220,0 210,224,266,280,9 71,240,99,268,2 37,343,93,399,0 218,59,302,143,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000099.jpg 176,369,190,383,0 30,116,44,130,5 46,323,102,379,9 310,159,338,187,1 19,171,47,199,2 80,241,108,269,6 101,151,213,263,6 132,70,216,154,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000100.jpg 244,233,266,255,9 132,4,146,18,3 27,41,49,63,9 258,38,300,80,6 318,276,360,318,0 225,204,253,232,9 31,210,59,238,9 310,51,352,93,3 83,13,111,41,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000101.jpg 287,176,329,218,8 328,39,384,95,6 61,200,117,256,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000102.jpg 6,68,20,82,6 197,307,211,321,0 196,348,238,390,8 76,101,104,129,1 249,118,305,174,7 184,235,212,263,8 16,307,44,335,1 83,187,167,271,1 274,182,386,294,0 186,40,270,124,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000103.jpg 108,15,122,29,5 29,165,51,187,3 303,192,317,206,3 255,189,297,231,6 126,149,168,191,0 188,157,244,213,5 155,51,211,107,1 156,335,212,391,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000104.jpg 377,206,399,228,3 137,53,151,67,4 343,361,365,383,7 186,247,228,289,6 218,168,260,210,4 201,17,313,129,3 19,147,131,259,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000105.jpg 71,9,93,31,2 167,78,181,92,7 163,273,185,295,1 351,142,379,170,3 14,131,70,187,9 69,217,97,245,9 288,25,344,81,1 38,285,66,313,1 224,207,266,249,2 73,53,185,165,1 195,298,307,410,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000106.jpg 329,206,413,290,5 184,4,296,116,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000107.jpg 39,393,53,407,1 219,95,241,117,5 184,333,206,355,0 178,135,220,177,4 60,162,88,190,9 14,244,98,328,0 223,294,335,406,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000108.jpg 71,351,85,365,8 347,260,403,316,6 248,243,304,299,2 299,65,355,121,1 41,23,97,79,3 139,271,167,299,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000109.jpg 374,375,388,389,6 254,160,276,182,9 77,305,91,319,2 309,314,351,356,5 202,348,230,376,3 205,230,317,342,2 208,30,292,114,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000110.jpg 294,26,308,40,6 234,269,248,283,1 382,29,404,51,7 10,9,38,37,8 248,207,290,249,6 127,13,155,41,1 171,87,199,115,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000111.jpg 37,34,93,90,0 198,113,240,155,1 346,291,402,347,3 256,135,284,163,4 69,178,97,206,8 117,330,173,386,0 232,164,344,276,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000112.jpg 99,294,141,336,1 22,383,50,411,7 235,152,291,208,7 198,34,226,62,0 3,131,45,173,0 134,202,246,314,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000113.jpg 120,152,134,166,7 21,335,43,357,1 283,340,297,354,6 25,241,67,283,8 299,372,327,400,3 179,302,235,358,5 92,112,204,224,5 238,14,350,126,3 41,292,153,404,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000114.jpg 127,282,141,296,3 183,27,205,49,0 55,38,69,52,1 65,177,149,261,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000115.jpg 292,352,320,380,4 80,336,108,364,0 321,37,405,121,2 58,61,142,145,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000116.jpg 204,394,218,408,4 359,325,381,347,8 156,92,178,114,9 46,19,102,75,9 198,307,254,363,6 312,333,340,361,6 110,288,138,316,3 310,264,366,320,8 234,44,346,156,4 71,157,183,269,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000117.jpg 361,214,403,256,4 254,335,310,391,7 68,341,124,397,7 134,37,190,93,3 204,262,260,318,1 109,144,193,228,7 310,52,394,136,2 208,129,292,213,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000118.jpg 376,78,390,92,5 29,147,43,161,4 254,130,276,152,8 195,136,251,192,3 363,321,391,349,0 300,70,328,98,0 195,364,237,406,6 11,25,123,137,9 71,181,155,265,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000119.jpg 143,274,165,296,2 98,2,120,24,6 247,312,275,340,4 264,158,306,200,4 114,153,170,209,2 155,95,197,137,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000120.jpg 175,223,189,237,5 99,12,141,54,9 354,25,410,81,4 137,128,221,212,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000121.jpg 78,310,92,324,6 190,369,204,383,6 37,19,93,75,7 254,358,296,400,9 128,73,156,101,4 108,304,150,346,6 150,16,178,44,9 247,10,359,122,3 0,88,84,172,2 23,179,107,263,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000122.jpg 360,36,402,78,0 365,175,407,217,3 84,30,126,72,6 383,134,411,162,2 263,282,291,310,5 229,41,271,83,2 30,194,114,278,1 230,142,342,254,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000123.jpg 346,351,360,365,8 269,145,283,159,4 305,60,319,74,6 281,353,309,381,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000124.jpg 169,87,197,115,6 322,207,364,249,7 27,135,55,163,2 352,334,394,376,5 80,152,164,236,9 248,78,360,190,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000125.jpg 250,297,272,319,3 275,312,289,326,8 237,326,251,340,6 184,234,240,290,5 359,163,415,219,5 168,100,224,156,8 286,2,342,58,3 46,259,158,371,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000126.jpg 347,168,361,182,5 11,162,25,176,7 26,92,68,134,6 122,98,150,126,2 154,100,210,156,8 183,250,239,306,3 51,295,93,337,9 25,196,109,280,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000127.jpg 17,172,31,186,1 345,194,387,236,2 154,139,196,181,2 324,252,366,294,2 219,150,303,234,1 109,6,193,90,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000128.jpg 328,311,342,325,9 50,284,106,340,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000129.jpg 99,273,113,287,9 236,320,264,348,4 290,50,346,106,3 104,74,160,130,4 156,61,268,173,0 107,230,219,342,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000130.jpg 27,114,41,128,0 377,57,391,71,8 160,91,182,113,2 254,166,296,208,7 262,12,290,40,3 147,37,189,79,3 262,238,290,266,0 55,36,139,120,4 296,48,408,160,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000131.jpg 58,359,100,401,9 93,143,121,171,8 128,303,240,415,7 262,162,374,274,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000132.jpg 238,136,252,150,5 305,3,333,31,0 279,243,335,299,6 3,229,59,285,3 283,194,311,222,6 364,156,406,198,2 122,143,234,255,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000133.jpg 260,75,302,117,6 360,35,388,63,7 73,29,157,113,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000134.jpg 68,182,96,210,5 327,311,411,395,1 216,225,300,309,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000135.jpg 66,108,88,130,7 280,378,302,400,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000136.jpg 305,360,327,382,6 150,368,172,390,8 364,114,406,156,2 170,365,212,407,6 274,181,316,223,7 166,253,208,295,3 353,246,395,288,0 316,49,358,91,1 54,62,138,146,0 17,157,129,269,4 129,169,213,253,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000137.jpg 104,361,132,389,0 110,318,138,346,1 291,188,319,216,3 23,321,79,377,2 180,3,208,31,8 198,88,310,200,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000138.jpg 37,258,59,280,9 153,64,209,120,3 197,216,225,244,3 305,364,347,406,9 124,260,166,302,6 231,79,343,191,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000139.jpg 183,58,205,80,6 212,265,226,279,9 88,82,144,138,6 322,255,406,339,8 268,20,352,104,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000140.jpg 381,63,395,77,1 94,69,116,91,9 71,257,93,279,2 264,4,320,60,0 2,195,44,237,4 108,111,150,153,4 252,223,280,251,0 174,33,230,89,5 276,75,318,117,4 287,162,399,274,7 282,310,366,394,3 134,287,246,399,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000141.jpg 94,69,122,97,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000142.jpg 302,320,316,334,7 247,346,289,388,7 211,215,267,271,3 143,241,171,269,0 241,50,325,134,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000143.jpg 161,93,175,107,1 41,379,55,393,4 130,257,172,299,6 240,211,268,239,3 107,67,135,95,5 37,247,65,275,7 260,247,288,275,0 4,63,116,175,2 142,170,226,254,4 227,288,311,372,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000144.jpg 221,147,243,169,8 120,200,162,242,0 96,260,124,288,5 272,296,384,408,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000145.jpg 149,53,177,81,3 237,231,279,273,5 55,74,83,102,5 178,322,206,350,2 161,251,203,293,8 310,289,394,373,4 9,176,121,288,0 190,116,302,228,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000146.jpg 208,215,230,237,5 191,66,205,80,0 389,118,403,132,0 218,184,246,212,7 333,368,375,410,3 276,140,360,224,9 261,256,345,340,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000147.jpg 14,304,70,360,1 120,56,162,98,1 155,198,239,282,1 288,209,400,321,4 225,58,337,170,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000148.jpg 248,165,276,193,3 121,101,163,143,3 344,108,400,164,1 22,10,64,52,4 205,179,233,207,5 299,3,341,45,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000149.jpg 127,278,141,292,7 184,70,240,126,1 194,226,236,268,0 278,123,320,165,5 38,30,150,142,6 152,284,264,396,4 302,25,414,137,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000150.jpg 273,229,287,243,7 194,227,216,249,7 203,317,245,359,6 74,110,186,222,1 314,8,398,92,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000151.jpg 355,168,411,224,7 309,213,351,255,7 134,68,218,152,0 222,144,306,228,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000152.jpg 318,182,332,196,7 310,109,332,131,2 352,141,394,183,0 165,17,207,59,0 12,275,68,331,9 90,164,146,220,8 140,300,224,384,0 35,76,119,160,9 163,71,247,155,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000153.jpg 24,229,46,251,0 371,282,385,296,6 33,135,61,163,4 326,247,354,275,9 3,3,87,87,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000154.jpg 337,180,359,202,0 190,240,212,262,7 129,192,143,206,1 36,256,92,312,5 167,294,223,350,0 32,206,74,248,3 215,154,327,266,9 139,34,223,118,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000155.jpg 18,273,40,295,1 118,308,132,322,5 340,85,368,113,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000156.jpg 265,340,287,362,3 97,16,111,30,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000157.jpg 210,157,238,185,7 44,144,72,172,1 60,291,144,375,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000158.jpg 175,335,189,349,7 92,95,114,117,5 208,32,236,60,7 327,249,383,305,4 71,252,127,308,8 229,47,313,131,8 107,126,219,238,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000159.jpg 123,106,151,134,0 346,327,388,369,4 92,257,176,341,2 15,41,99,125,9 177,105,289,217,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000160.jpg 68,157,124,213,7 288,81,316,109,6 261,192,303,234,1 358,39,386,67,9 147,202,203,258,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000161.jpg 60,70,74,84,3 135,357,149,371,2 22,96,44,118,0 309,254,351,296,3 18,359,74,415,8 185,0,213,28,8 168,151,224,207,1 249,29,291,71,0 57,94,141,178,4 169,214,281,326,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000162.jpg 28,259,84,315,7 202,73,244,115,6 66,219,108,261,4 320,195,376,251,1 61,25,89,53,8 187,292,229,334,1 17,140,101,224,2 142,136,254,248,6 256,266,368,378,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000163.jpg 67,125,89,147,1 118,308,132,322,3 313,379,327,393,5 281,226,323,268,8 93,132,135,174,8 18,298,130,410,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000164.jpg 226,2,338,114,7 87,50,199,162,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000165.jpg 317,304,331,318,1 51,36,163,148,3 160,118,244,202,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000166.jpg 385,122,407,144,2 396,376,410,390,8 27,3,69,45,8 347,94,375,122,3 146,243,174,271,9 207,278,235,306,8 105,100,189,184,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000167.jpg 141,203,163,225,6 220,368,242,390,8 308,293,336,321,2 222,329,250,357,1 349,258,391,300,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000168.jpg 295,135,317,157,2 217,91,231,105,4 223,252,265,294,0 372,130,400,158,5 127,371,169,413,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000169.jpg 358,77,372,91,0 80,10,102,32,1 263,31,277,45,3 309,244,365,300,9 322,88,378,144,6 184,246,226,288,4 130,212,172,254,2 238,102,294,158,5 38,76,150,188,6 83,270,195,382,3 259,300,371,412,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000170.jpg 120,299,142,321,8 50,106,162,218,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000171.jpg 33,16,55,38,2 86,222,100,236,3 312,349,334,371,6 302,181,358,237,1 249,364,291,406,7 338,247,394,303,9 203,31,259,87,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000172.jpg 317,398,331,412,0 19,71,41,93,7 153,86,209,142,6 129,209,157,237,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000173.jpg 317,206,373,262,7 33,53,89,109,9 294,387,322,415,8 128,260,240,372,8 111,95,195,179,8 211,82,323,194,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000174.jpg 287,344,301,358,1 265,103,287,125,6 139,26,153,40,1 353,196,409,252,2 181,148,237,204,7 196,243,252,299,0 268,1,310,43,3 103,293,215,405,8 258,264,342,348,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000175.jpg 239,23,295,79,8 6,123,48,165,3 237,221,321,305,0 330,29,414,113,2 129,88,241,200,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000176.jpg 162,142,204,184,9 89,110,131,152,5 232,140,316,224,3 305,284,389,368,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000177.jpg 218,240,240,262,8 84,121,98,135,8 24,267,66,309,5 145,290,187,332,1 170,352,212,394,4 248,94,290,136,7 47,226,89,268,3 61,172,103,214,0 124,198,208,282,4 208,290,320,402,6 256,170,368,282,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000178.jpg 216,278,272,334,8 298,36,340,78,1 264,324,320,380,0 0,40,28,68,3 136,297,164,325,4 61,162,173,274,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000179.jpg 394,311,408,325,9 129,124,143,138,1 272,156,300,184,8 180,327,222,369,9 267,351,295,379,3 320,40,376,96,3 174,19,230,75,7 125,228,209,312,6 134,123,218,207,6 0,29,112,141,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000180.jpg 168,140,190,162,4 327,387,341,401,9 9,330,65,386,4 272,123,328,179,5 136,71,192,127,4 158,324,214,380,4 34,13,76,55,8 328,312,384,368,0 202,17,314,129,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000181.jpg 157,137,213,193,7 119,201,161,243,1 0,77,56,133,3 275,117,317,159,0 30,126,114,210,9 150,275,262,387,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000182.jpg 270,145,292,167,5 299,76,313,90,0 105,308,127,330,6 322,269,364,311,6 185,27,213,55,9 249,34,277,62,9 298,195,340,237,2 104,149,146,191,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000183.jpg 38,46,52,60,4 301,208,329,236,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000184.jpg 124,310,208,394,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000185.jpg 239,30,253,44,7 287,22,309,44,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000186.jpg 325,299,353,327,0 188,15,230,57,6 34,223,76,265,0 241,262,297,318,0 204,59,246,101,5 58,295,86,323,4 278,130,390,242,0 36,127,120,211,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000187.jpg 107,186,121,200,1 206,19,248,61,5 283,125,339,181,4 34,34,118,118,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000188.jpg 22,136,44,158,6 115,130,129,144,2 24,255,66,297,9 152,55,264,167,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000189.jpg 165,23,179,37,8 294,143,316,165,7 225,307,281,363,6 282,22,324,64,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000190.jpg 128,265,150,287,5 19,163,61,205,8 328,1,412,85,7 145,269,229,353,8 261,159,345,243,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000191.jpg 45,325,59,339,9 216,138,244,166,9 70,13,98,41,4 242,237,284,279,8 252,140,280,168,8 354,305,396,347,7 95,328,137,370,6 67,240,151,324,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000192.jpg 326,319,340,333,6 213,297,235,319,0 257,233,279,255,3 5,300,33,328,4 101,103,185,187,6 269,248,381,360,2 47,284,159,396,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000193.jpg 192,180,206,194,7 38,205,66,233,3 371,221,413,263,4 137,118,193,174,7 279,237,335,293,0 223,182,279,238,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000194.jpg 215,359,229,373,4 27,346,41,360,0 160,36,174,50,3 21,242,63,284,0 60,173,144,257,7 239,224,323,308,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000195.jpg 268,160,282,174,1 386,37,400,51,5 81,127,95,141,4 216,326,272,382,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000196.jpg 318,8,346,36,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000197.jpg 326,236,348,258,0 281,310,295,324,1 46,117,88,159,8 104,101,216,213,3 58,229,170,341,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000198.jpg 310,232,352,274,3 232,236,260,264,5 73,269,115,311,4 172,258,214,300,7 70,148,182,260,3 258,7,370,119,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000199.jpg 374,209,396,231,7 253,30,267,44,1 373,139,395,161,1 204,110,260,166,2 76,147,160,231,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000200.jpg 35,40,57,62,1 100,187,122,209,1 263,257,291,285,7 106,266,162,322,9 310,289,352,331,2 168,201,224,257,4 260,9,302,51,1 131,30,159,58,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000201.jpg 91,193,113,215,5 232,62,260,90,4 311,85,395,169,7 126,186,210,270,8 254,274,366,386,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000202.jpg 215,327,237,349,2 170,345,198,373,9 76,243,132,299,2 292,105,348,161,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000203.jpg 375,340,389,354,4 37,136,51,150,8 95,91,137,133,0 340,56,368,84,7 247,191,303,247,3 126,165,168,207,4 205,1,233,29,8 300,100,412,212,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000204.jpg 357,95,371,109,4 119,350,147,378,5 240,250,268,278,0 242,367,284,409,1 180,349,208,377,8 317,114,359,156,0 161,238,245,322,2 109,115,193,199,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000205.jpg 278,300,300,322,6 51,279,73,301,0 301,111,343,153,4 90,253,202,365,3 155,27,267,139,7 52,71,136,155,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000206.jpg 199,323,213,337,3 309,257,365,313,7 346,178,374,206,2 211,98,295,182,0 331,67,415,151,8 75,245,187,357,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000207.jpg 368,290,382,304,6 1,2,15,16,7 279,231,293,245,7 266,201,294,229,9 287,84,315,112,7 335,229,377,271,6 25,13,53,41,2 167,271,209,313,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000208.jpg 103,148,125,170,3 100,292,122,314,8 145,198,159,212,4 293,248,335,290,6 304,295,360,351,7 239,70,295,126,1 257,138,369,250,4 120,5,232,117,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000209.jpg 36,318,50,332,4 140,15,162,37,7 144,331,200,387,1 284,227,326,269,9 243,88,271,116,4 287,40,315,68,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000210.jpg 186,226,208,248,7 111,114,133,136,5 295,124,337,166,6 111,276,153,318,0 250,198,306,254,4 0,386,28,414,4 320,195,404,279,6 70,161,182,273,6 273,284,385,396,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000211.jpg 66,202,88,224,1 47,373,69,395,5 147,244,189,286,0 21,133,77,189,0 187,322,215,350,4 168,107,252,191,2 239,250,351,362,3 264,143,376,255,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000212.jpg 287,318,309,340,4 55,293,77,315,0 27,174,41,188,1 350,301,392,343,9 325,180,353,208,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000213.jpg 318,165,340,187,7 163,86,219,142,7 171,181,283,293,5 245,323,329,407,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000214.jpg 172,34,186,48,4 313,175,355,217,9 287,111,329,153,7 200,234,256,290,0 169,107,197,135,2 317,278,373,334,8 20,199,132,311,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000215.jpg 322,198,364,240,7 226,1,268,43,5 358,8,386,36,2 171,116,283,228,6 39,70,151,182,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000216.jpg 208,45,222,59,0 232,317,254,339,5 377,168,391,182,9 170,358,198,386,3 89,216,117,244,2 255,159,311,215,8 162,262,218,318,2 264,332,320,388,1 325,174,353,202,1 148,101,260,213,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000217.jpg 225,141,337,253,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000218.jpg 287,255,301,269,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000219.jpg 313,15,335,37,5 335,100,357,122,7 139,253,195,309,0 35,225,77,267,7 181,157,293,269,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000220.jpg 134,218,148,232,3 200,319,242,361,5 92,291,120,319,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000221.jpg 347,401,361,415,3 70,192,98,220,0 381,135,409,163,2 123,226,165,268,7 90,229,118,257,8 338,86,380,128,9 35,290,147,402,2 219,187,303,271,3 44,78,128,162,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000222.jpg 84,107,98,121,5 393,250,415,272,4 20,325,76,381,3 191,124,219,152,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000223.jpg 344,93,358,107,8 19,189,61,231,2 43,53,127,137,5 234,113,318,197,8 1,290,113,402,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000224.jpg 275,98,289,112,0 169,24,183,38,3 222,35,278,91,6 216,101,244,129,8 333,6,375,48,3 266,134,378,246,0 52,210,164,322,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000225.jpg 228,208,250,230,8 274,373,302,401,1 255,291,311,347,2 264,208,320,264,1 1,186,85,270,5 287,49,399,161,6 104,162,216,274,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000226.jpg 342,214,364,236,9 158,43,172,57,4 121,36,143,58,4 227,126,269,168,7 260,36,302,78,7 307,295,363,351,6 155,155,197,197,7 50,276,106,332,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000227.jpg 283,210,311,238,2 280,299,308,327,9 237,317,279,359,0 141,126,197,182,2 279,151,321,193,2 103,224,215,336,8 193,45,305,157,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000228.jpg 145,208,201,264,0 254,16,296,58,6 13,136,125,248,2 232,247,344,359,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000229.jpg 111,26,133,48,4 280,312,302,334,8 317,387,339,409,9 243,115,285,157,8 29,173,85,229,0 17,37,73,93,4 318,180,346,208,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000230.jpg 308,389,322,403,4 122,190,136,204,4 153,217,209,273,0 242,138,270,166,9 58,29,86,57,2 13,239,125,351,0 189,33,301,145,3 303,125,415,237,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000231.jpg 175,106,197,128,3 307,124,321,138,4 161,328,183,350,5 73,54,101,82,3 102,202,214,314,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000232.jpg 377,0,391,14,4 127,236,141,250,9 65,324,79,338,4 174,123,216,165,6 159,95,187,123,5 387,34,415,62,1 251,275,363,387,5 278,71,362,155,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000233.jpg 363,80,405,122,3 72,309,156,393,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000234.jpg 67,362,95,390,0 16,221,72,277,5 271,80,299,108,1 104,116,146,158,4 291,82,403,194,6 165,69,277,181,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000235.jpg 151,114,165,128,4 351,372,373,394,7 286,355,342,411,0 8,275,50,317,1 7,127,119,239,6 155,28,267,140,0 197,258,281,342,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000236.jpg 338,61,380,103,5 292,354,334,396,8 19,158,75,214,6 210,355,238,383,7 26,339,82,395,9 154,3,210,59,3 189,71,273,155,1 123,150,235,262,9 31,245,115,329,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000237.jpg 370,204,398,232,9 345,59,387,101,5 10,111,66,167,4 335,363,377,405,0 206,236,262,292,6 81,291,109,319,6 205,19,317,131,4 55,44,139,128,4 113,239,197,323,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000238.jpg 47,80,61,94,8 328,30,370,72,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000239.jpg 236,287,258,309,8 152,145,166,159,2 161,289,175,303,4 19,30,47,58,4 288,326,372,410,1 197,156,281,240,2 110,288,222,400,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000240.jpg 57,244,71,258,6 225,222,239,236,1 176,240,190,254,7 7,368,49,410,8 289,197,401,309,4 6,70,90,154,8 56,290,168,402,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000241.jpg 303,352,325,374,0 101,189,143,231,1 340,75,396,131,8 164,90,220,146,1 155,268,197,310,8 360,326,402,368,5 6,13,90,97,3 119,328,203,412,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000242.jpg 144,57,166,79,7 104,91,126,113,8 134,90,176,132,0 21,307,77,363,2 40,240,82,282,0 320,43,348,71,8 272,345,300,373,1 258,60,286,88,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000243.jpg 168,340,190,362,8 195,291,209,305,5 70,247,92,269,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000244.jpg 165,287,187,309,8 234,305,248,319,1 46,377,60,391,9 56,313,112,369,8 294,161,350,217,9 241,177,297,233,4 1,200,85,284,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000245.jpg 313,284,327,298,9 142,284,198,340,1 198,237,226,265,6 84,244,126,286,3 127,228,155,256,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000246.jpg 199,1,241,43,3 87,285,143,341,2 98,368,126,396,5 144,287,186,329,3 301,177,329,205,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000247.jpg 208,207,222,221,6 203,134,217,148,0 100,359,142,401,7 194,76,250,132,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000248.jpg 319,313,333,327,0 295,11,337,53,5 215,55,243,83,1 274,138,386,250,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000249.jpg 10,162,38,190,2 69,76,125,132,9 174,4,202,32,5 128,20,170,62,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000250.jpg 268,10,324,66,5 229,322,285,378,3 60,224,88,252,4 318,157,402,241,4 80,125,192,237,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000251.jpg 209,220,265,276,9 27,196,83,252,8 111,216,153,258,1 27,61,69,103,9 143,258,171,286,5 73,379,101,407,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000252.jpg 161,283,175,297,8 168,210,182,224,2 51,246,79,274,1 136,367,178,409,2 5,146,33,174,8 174,35,286,147,2 281,284,393,396,3 67,154,151,238,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000253.jpg 44,34,66,56,3 297,301,353,357,1 340,70,382,112,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000254.jpg 247,382,275,410,9 192,113,248,169,5 190,232,232,274,9 324,42,352,70,2 54,7,110,63,2 93,150,121,178,1 42,224,154,336,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000255.jpg 71,98,93,120,3 214,36,236,58,4 41,190,125,274,7 243,67,327,151,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000256.jpg 39,332,95,388,5 213,32,255,74,6 273,315,301,343,6 319,15,347,43,7 29,39,141,151,5 270,211,382,323,8 96,250,208,362,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000257.jpg 306,146,362,202,8 144,164,172,192,6 165,360,207,402,8 32,238,88,294,8 259,282,371,394,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000258.jpg 184,259,206,281,0 313,236,341,264,5 219,8,247,36,1 93,31,135,73,8 264,240,306,282,9 112,93,168,149,8 189,164,245,220,4 69,309,153,393,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000259.jpg 200,118,214,132,7 106,371,128,393,5 205,310,219,324,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000260.jpg 52,307,108,363,0 195,264,307,376,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000261.jpg 77,379,91,393,9 401,144,415,158,3 276,17,304,45,4 118,257,174,313,2 273,70,315,112,2 171,106,199,134,0 262,175,374,287,3 5,302,89,386,9 201,270,285,354,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000262.jpg 317,46,339,68,0 385,147,407,169,9 123,91,165,133,6 139,250,167,278,8 68,104,124,160,1 165,31,221,87,1 264,365,306,407,0 151,132,179,160,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000263.jpg 297,186,319,208,6 283,365,297,379,8 115,366,137,388,1 237,7,293,63,3 169,193,281,305,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000264.jpg 126,80,168,122,1 41,50,69,78,7 368,322,396,350,1 8,245,64,301,5 201,214,257,270,8 238,286,350,398,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000265.jpg 34,217,56,239,7 355,152,369,166,0 350,94,364,108,0 183,143,239,199,1 29,305,85,361,0 20,25,132,137,9 310,8,394,92,4 56,171,140,255,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000266.jpg 296,211,338,253,5 14,232,56,274,2 235,181,263,209,4 248,6,304,62,9 326,260,382,316,7 172,145,214,187,8 130,216,242,328,8 4,105,116,217,4 111,24,223,136,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000267.jpg 180,115,194,129,3 182,21,204,43,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000268.jpg 203,263,231,291,6 101,121,143,163,0 104,60,146,102,5 234,198,346,310,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000269.jpg 108,62,130,84,1 68,342,90,364,7 106,356,120,370,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000270.jpg 354,150,410,206,6 241,38,297,94,9 11,7,67,63,4 84,90,196,202,4 147,219,231,303,8 265,278,377,390,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000271.jpg 142,385,156,399,9 249,347,263,361,4 204,229,218,243,6 249,97,291,139,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000272.jpg 296,143,338,185,7 222,11,334,123,4 148,297,232,381,1 199,118,283,202,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000273.jpg 30,167,52,189,2 100,326,122,348,6 380,356,408,384,8 149,87,233,171,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000274.jpg 41,313,63,335,8 75,125,89,139,8 61,162,117,218,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000275.jpg 138,75,152,89,7 94,333,116,355,1 80,3,102,25,6 190,193,302,305,7 121,106,205,190,8 256,110,340,194,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000276.jpg 205,248,247,290,9 111,69,139,97,1 187,97,243,153,9 179,299,291,411,1 256,59,368,171,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000277.jpg 28,59,56,87,3 326,208,368,250,3 353,61,381,89,7 116,17,158,59,4 216,149,258,191,9 37,167,149,279,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000278.jpg 289,366,311,388,9 35,7,57,29,9 142,57,164,79,7 29,29,71,71,6 323,177,379,233,2 279,60,363,144,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000279.jpg 252,277,274,299,9 84,31,196,143,4 63,217,175,329,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000280.jpg 212,316,226,330,8 337,85,351,99,3 4,208,18,222,1 242,355,270,383,3 71,254,113,296,2 194,347,236,389,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000281.jpg 171,382,193,404,5 341,378,363,400,6 151,139,173,161,3 63,153,105,195,1 128,336,170,378,1 33,108,75,150,0 153,65,209,121,9 141,0,197,56,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000282.jpg 298,385,312,399,8 298,105,340,147,9 329,360,371,402,8 190,191,218,219,1 317,325,345,353,1 184,150,226,192,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000283.jpg 258,372,300,414,6 132,209,160,237,9 81,138,123,180,1 92,246,148,302,5 219,250,275,306,7 359,154,387,182,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000284.jpg 228,207,340,319,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000285.jpg 87,6,109,28,3 123,82,145,104,0 165,372,193,400,9 270,258,298,286,4 86,175,114,203,5 292,104,320,132,6 189,207,273,291,0 140,31,224,115,1 42,209,126,293,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000286.jpg 227,375,249,397,6 47,301,103,357,3 145,81,201,137,8 29,11,85,67,2 314,308,370,364,2 241,282,297,338,6 201,241,229,269,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000287.jpg 317,358,331,372,7 14,109,70,165,4 93,241,149,297,6 120,188,176,244,7 69,323,153,407,7 238,302,322,386,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000288.jpg 264,396,278,410,1 32,23,46,37,1 105,366,119,380,2 170,111,226,167,5 147,284,203,340,2 172,12,200,40,1 142,171,226,255,2 322,6,406,90,6 2,157,114,269,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000289.jpg 50,45,64,59,7 206,331,228,353,0 319,182,361,224,0 135,19,191,75,2 163,199,219,255,4 252,169,308,225,4 94,83,136,125,6 296,12,352,68,8 44,183,156,295,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000290.jpg 274,327,288,341,3 36,334,64,362,1 232,149,260,177,9 267,67,323,123,5 25,241,67,283,5 99,107,211,219,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000291.jpg 190,35,246,91,5 123,175,165,217,7 249,278,361,390,7 317,180,401,264,8 206,149,290,233,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000292.jpg 202,169,244,211,4 180,264,236,320,5 61,356,89,384,4 55,131,111,187,2 139,168,167,196,6 296,167,324,195,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000293.jpg 85,335,99,349,1 54,351,110,407,9 270,41,298,69,9 93,248,121,276,8 182,202,224,244,8 317,43,401,127,5 207,85,291,169,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000294.jpg 166,204,180,218,5 228,350,250,372,7 293,234,307,248,7 8,186,64,242,1 335,44,377,86,7 236,203,348,315,0 5,246,117,358,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000295.jpg 258,19,272,33,3 282,248,324,290,3 138,236,194,292,0 29,263,57,291,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000296.jpg 350,265,372,287,6 194,11,216,33,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000297.jpg 346,259,360,273,7 119,249,141,271,3 243,319,285,361,2 235,112,291,168,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000298.jpg 123,316,137,330,3 176,297,260,381,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000299.jpg 75,387,97,409,6 40,290,96,346,0 1,104,57,160,7 360,22,388,50,0 185,94,227,136,6 25,247,67,289,8 16,15,100,99,0 173,155,285,267,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000300.jpg 311,132,333,154,6 111,354,125,368,1 111,187,139,215,9 286,335,328,377,3 302,184,358,240,4 18,292,60,334,9 189,165,231,207,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000301.jpg 202,127,224,149,6 349,66,371,88,3 87,259,109,281,7 130,108,172,150,1 278,164,306,192,8 272,204,300,232,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000302.jpg 1,139,23,161,8 75,359,117,401,9 89,18,201,130,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000303.jpg 160,82,216,138,5 78,36,134,92,1 30,45,58,73,1 38,16,66,44,4 261,216,303,258,5 112,284,196,368,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000304.jpg 133,140,175,182,8 229,168,257,196,0 326,196,410,280,9 11,82,123,194,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000305.jpg 298,286,312,300,4 304,192,318,206,3 271,156,285,170,3 3,97,31,125,6 55,256,111,312,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000306.jpg 303,90,325,112,7 140,87,154,101,7 187,184,215,212,4 143,278,185,320,9 315,29,343,57,1 98,20,126,48,1 154,141,196,183,2 254,160,338,244,0 193,313,277,397,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000307.jpg 117,342,139,364,5 174,356,188,370,8 302,56,324,78,3 81,27,165,111,8 145,183,257,295,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000308.jpg 364,272,378,286,2 67,101,109,143,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000309.jpg 70,311,92,333,8 19,192,41,214,9 327,144,341,158,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000310.jpg 316,345,330,359,0 268,177,282,191,3 36,227,78,269,7 183,232,239,288,7 112,382,140,410,7 15,300,43,328,4 251,121,363,233,6 145,296,257,408,3 93,215,177,299,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000311.jpg 393,139,415,161,0 64,89,78,103,3 159,305,173,319,7 322,41,350,69,1 71,338,127,394,4 133,135,175,177,1 295,329,323,357,4 373,263,401,291,2 95,219,207,331,5 210,117,294,201,6 11,32,123,144,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000312.jpg 204,3,218,17,4 243,367,257,381,2 268,355,324,411,3 184,35,240,91,8 150,145,206,201,5 370,25,398,53,2 358,286,386,314,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000313.jpg 269,255,291,277,4 132,63,146,77,7 43,85,57,99,4 340,81,382,123,3 129,177,171,219,2 200,290,228,318,9 251,326,307,382,7 0,356,56,412,4 89,189,131,231,3 91,45,203,157,4 237,102,349,214,1 282,225,394,337,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000314.jpg 384,152,398,166,2 205,40,219,54,9 330,16,344,30,9 272,44,300,72,2 195,154,251,210,0 231,20,273,62,6 102,44,158,100,1 222,329,278,385,5 61,266,145,350,0 274,135,358,219,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000315.jpg 379,336,401,358,6 65,9,79,23,0 128,139,240,251,1 172,248,284,360,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000316.jpg 142,184,164,206,7 349,107,391,149,9 102,348,130,376,3 205,178,261,234,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000317.jpg 269,233,291,255,6 67,293,109,335,0 339,332,395,388,4 137,215,165,243,8 116,280,144,308,2 230,154,286,210,7 214,325,298,409,7 294,200,378,284,2 96,34,208,146,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000318.jpg 222,372,264,414,5 320,319,362,361,7 164,309,206,351,3 108,95,150,137,0 217,145,329,257,8 106,146,218,258,6 209,274,293,358,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000319.jpg 174,21,196,43,1 277,300,305,328,5 278,369,320,411,6 175,155,203,183,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000320.jpg 224,371,266,413,8 371,306,399,334,0 270,233,312,275,9 286,313,328,355,1 84,159,140,215,7 288,37,400,149,0 38,33,122,117,0 113,273,225,385,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000321.jpg 196,136,224,164,7 34,222,62,250,1 231,250,343,362,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000322.jpg 139,165,161,187,4 292,20,320,48,8 193,176,305,288,1 63,242,147,326,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000323.jpg 348,28,370,50,9 134,43,156,65,4 163,326,191,354,6 278,286,334,342,9 86,227,142,283,6 292,44,334,86,5 382,120,410,148,0 74,33,116,75,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000324.jpg 130,22,152,44,8 336,116,358,138,1 164,85,178,99,1 302,246,358,302,5 214,78,242,106,4 122,230,150,258,1 51,237,79,265,6 176,155,288,267,9 68,79,152,163,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000325.jpg 272,39,328,95,1 128,312,184,368,5 117,10,145,38,0 186,5,214,33,3 24,355,66,397,7 64,339,120,395,2 3,4,115,116,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000326.jpg 175,147,197,169,9 245,116,267,138,6 111,324,125,338,4 119,39,175,95,1 14,79,56,121,5 309,317,351,359,3 222,254,264,296,1 171,263,199,291,9 76,210,160,294,3 289,72,401,184,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000327.jpg 21,123,105,207,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000328.jpg 82,185,104,207,1 107,232,129,254,4 26,119,48,141,2 338,318,394,374,9 107,169,163,225,3 189,126,217,154,4 145,347,201,403,8 211,131,323,243,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000329.jpg 36,187,148,299,4 28,57,140,169,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000330.jpg 278,7,300,29,4 31,21,45,35,7 162,250,176,264,5 258,197,370,309,5 25,254,137,366,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000331.jpg 164,10,186,32,1 364,39,406,81,8 4,262,88,346,3 99,64,183,148,6 325,304,409,388,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000332.jpg 105,248,127,270,5 119,302,141,324,2 295,274,309,288,2 153,27,195,69,4 352,113,394,155,1 171,282,199,310,1 15,109,127,221,3 152,130,264,242,1 237,299,349,411,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000333.jpg 313,138,341,166,9 80,1,122,43,5 163,90,191,118,8 235,68,277,110,2 261,176,303,218,9 248,299,276,327,4 67,239,179,351,0 303,264,415,376,0 53,102,165,214,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000334.jpg 12,188,34,210,1 370,309,384,323,7 6,226,90,310,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000335.jpg 393,352,407,366,7 340,227,382,269,7 89,23,131,65,0 258,289,370,401,3 50,179,162,291,0 179,8,263,92,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000336.jpg 166,225,180,239,7 396,100,410,114,2 12,288,34,310,7 248,321,304,377,9 33,100,75,142,4 310,274,352,316,5 117,324,145,352,0 232,202,288,258,1 111,187,223,299,3 172,10,284,122,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000337.jpg 398,240,412,254,0 143,40,157,54,9 284,208,340,264,4 223,142,251,170,3 33,280,75,322,9 259,377,287,405,1 149,367,177,395,4 226,31,338,143,3 64,138,176,250,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000338.jpg 230,52,286,108,2 231,349,287,405,2 33,169,145,281,0 154,242,266,354,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000339.jpg 24,97,52,125,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000340.jpg 208,333,230,355,1 241,154,255,168,7 285,5,307,27,8 198,192,254,248,0 50,196,106,252,7 316,238,358,280,0 95,329,179,413,8 311,78,395,162,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000341.jpg 337,181,351,195,1 29,243,43,257,6 59,297,87,325,9 297,141,325,169,7 159,354,201,396,5 233,209,261,237,5 157,189,213,245,2 0,0,112,112,7 197,256,309,368,3 163,43,275,155,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000342.jpg 290,373,318,401,0 274,199,316,241,6 82,139,110,167,9 152,263,208,319,9 160,187,216,243,0 50,265,134,349,1 150,18,262,130,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000343.jpg 75,172,89,186,2 198,354,212,368,6 312,287,354,329,3 219,263,247,291,7 202,21,244,63,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000344.jpg 294,50,308,64,3 54,322,68,336,1 321,155,343,177,2 37,125,79,167,6 196,27,308,139,4 143,143,227,227,5 89,234,201,346,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000345.jpg 117,346,145,374,0 253,99,309,155,8 152,237,236,321,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000346.jpg 385,373,399,387,8 176,139,218,181,2 96,286,152,342,2 206,259,262,315,0 215,126,327,238,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000347.jpg 38,150,52,164,7 52,355,74,377,4 197,77,225,105,7 177,384,205,412,5 324,167,352,195,7 355,213,411,269,1 293,291,349,347,3 153,326,195,368,0 109,133,221,245,2 25,41,109,125,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000348.jpg 188,325,202,339,9 86,350,108,372,9 105,211,127,233,2 273,46,385,158,4 69,99,153,183,7 227,168,311,252,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000349.jpg 146,281,168,303,6 126,57,140,71,0 27,343,41,357,7 176,247,218,289,0 223,8,335,120,0 30,54,114,138,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000350.jpg 166,159,188,181,9 38,135,52,149,8 356,221,378,243,0 77,36,105,64,9 194,171,236,213,7 2,91,114,203,5 206,80,290,164,7 52,215,164,327,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000351.jpg 159,304,181,326,5 27,203,83,259,4 97,29,125,57,9 345,65,373,93,1 109,299,137,327,3 264,56,306,98,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000352.jpg 53,122,67,136,7 286,20,328,62,6 70,25,126,81,0 91,200,175,284,5 76,297,188,409,9 202,71,286,155,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000353.jpg 11,372,25,386,7 293,266,315,288,2 350,115,406,171,3 134,177,246,289,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000354.jpg 83,203,105,225,1 233,249,247,263,6 147,67,169,89,8 333,68,361,96,1 311,114,367,170,3 194,156,278,240,1 54,242,166,354,2 50,44,134,128,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000355.jpg 79,167,107,195,6 347,321,389,363,6 98,9,126,37,4 38,266,150,378,4 307,135,391,219,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000356.jpg 344,129,358,143,2 240,253,262,275,8 167,238,181,252,0 94,263,150,319,7 201,58,243,100,5 63,247,91,275,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000357.jpg 91,104,147,160,5 252,164,364,276,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000358.jpg 81,51,123,93,9 176,231,232,287,6 331,348,359,376,8 166,33,194,61,3 313,54,369,110,8 344,291,400,347,4 255,172,367,284,6 21,120,133,232,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000359.jpg 249,380,271,402,6 6,132,34,160,2 209,334,237,362,5 172,199,228,255,1 121,62,205,146,2 31,229,143,341,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000360.jpg 60,376,88,404,8 356,275,412,331,6 188,227,244,283,2 313,10,397,94,0 81,248,165,332,3 9,74,121,186,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000361.jpg 304,157,318,171,8 194,307,216,329,8 340,35,362,57,3 343,290,385,332,8 372,39,414,81,6 122,270,178,326,0 164,187,192,215,8 353,165,381,193,2 338,99,394,155,4 129,83,241,195,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000362.jpg 200,107,222,129,6 266,64,280,78,3 78,269,100,291,9 215,251,327,363,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000363.jpg 350,45,364,59,7 224,226,246,248,0 112,273,134,295,8 324,171,352,199,9 244,245,356,357,5 206,20,318,132,5 31,292,143,404,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000364.jpg 89,127,111,149,7 187,121,229,163,2 175,295,217,337,4 250,346,278,374,5 14,130,56,172,2 9,371,51,413,3 218,206,274,262,4 302,188,414,300,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000365.jpg 248,124,290,166,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000366.jpg 212,280,254,322,0 271,205,299,233,1 33,119,145,231,4 72,257,156,341,2 268,60,352,144,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000367.jpg 242,219,284,261,3 278,4,320,46,7 344,139,386,181,0 110,98,194,182,8 93,184,205,296,3 217,258,329,370,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000368.jpg 51,291,73,313,9 255,37,269,51,1 298,274,354,330,6 52,87,108,143,6 247,66,359,178,6 85,327,169,411,4 166,165,278,277,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000369.jpg 193,343,215,365,1 284,148,306,170,8 315,21,371,77,2 97,269,125,297,2 235,247,277,289,1 39,227,67,255,5 295,372,337,414,3 306,88,362,144,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000370.jpg 279,234,301,256,6 180,95,236,151,8 285,271,313,299,2 140,154,182,196,6 196,201,238,243,5 328,199,356,227,1 159,286,271,398,4 50,295,162,407,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000371.jpg 20,236,42,258,6 133,389,155,411,4 88,138,102,152,0 54,254,110,310,6 378,96,406,124,5 266,202,378,314,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000372.jpg 392,273,406,287,2 55,244,111,300,3 161,269,189,297,0 386,61,414,89,2 82,134,110,162,0 141,20,253,132,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000373.jpg 325,339,353,367,2 289,85,331,127,1 323,372,351,400,2 152,87,264,199,6 160,310,244,394,7 39,162,151,274,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000374.jpg 184,355,206,377,2 33,154,55,176,1 271,297,285,311,1 344,185,400,241,9 170,95,198,123,8 230,204,258,232,8 312,37,354,79,0 186,242,228,284,1 9,238,93,322,8 274,89,358,173,0 117,157,201,241,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000375.jpg 317,32,331,46,5 219,132,241,154,7 62,256,76,270,0 255,91,297,133,0 197,342,225,370,0 193,231,277,315,3 62,299,146,383,3 266,156,350,240,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000376.jpg 313,371,335,393,4 282,124,324,166,3 162,323,190,351,5 47,347,103,403,1 44,14,72,42,8 60,105,172,217,7 233,168,345,280,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000377.jpg 166,349,180,363,1 260,90,316,146,9 98,46,154,102,5 124,153,208,237,0 138,296,250,408,2 330,308,414,392,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000378.jpg 5,225,19,239,3 16,253,58,295,0 282,324,324,366,5 271,233,299,261,8 135,271,191,327,7 197,12,225,40,6 316,264,358,306,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000379.jpg 297,246,319,268,1 88,171,110,193,8 225,6,267,48,9 159,107,215,163,9 387,306,415,334,4 82,356,124,398,1 97,33,125,61,3 264,69,376,181,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000380.jpg 374,156,396,178,8 79,66,121,108,9 9,143,65,199,2 123,193,207,277,1 95,292,207,404,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000381.jpg 371,256,385,270,3 278,255,390,367,8 199,0,283,84,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000382.jpg 270,224,292,246,2 341,359,355,373,4 274,26,288,40,2 312,233,354,275,2 175,235,231,291,9 189,96,231,138,5 346,28,374,56,3 75,317,131,373,5 273,104,357,188,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000383.jpg 207,245,221,259,3 176,257,198,279,7 83,41,105,63,8 159,106,187,134,2 121,34,177,90,9 251,151,293,193,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000384.jpg 191,15,247,71,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000385.jpg 354,121,376,143,8 336,319,392,375,8 141,346,197,402,0 31,374,59,402,2 110,140,152,182,4 200,240,284,324,8 39,193,151,305,9 16,5,128,117,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000386.jpg 126,350,140,364,9 200,388,214,402,1 353,116,395,158,4 271,229,299,257,8 185,20,297,132,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000387.jpg 188,246,210,268,1 12,174,26,188,4 19,394,33,408,0 1,251,29,279,1 303,297,345,339,0 360,268,388,296,1 176,298,232,354,1 52,189,108,245,7 98,41,182,125,6 280,32,364,116,7 16,299,128,411,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000388.jpg 311,264,325,278,5 357,143,379,165,5 211,25,267,81,6 55,116,111,172,0 124,232,166,274,7 121,169,149,197,9 168,296,252,380,6 183,117,267,201,6 257,162,369,274,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000389.jpg 305,62,361,118,3 19,287,47,315,8 88,147,172,231,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000390.jpg 111,149,125,163,4 328,386,342,400,2 180,159,194,173,4 45,280,87,322,8 356,164,412,220,2 94,222,122,250,3 193,232,277,316,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000391.jpg 379,219,393,233,6 389,116,403,130,1 115,273,129,287,1 126,257,182,313,1 334,259,376,301,3 211,77,267,133,3 171,158,227,214,1 354,52,410,108,4 37,109,121,193,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000392.jpg 164,240,186,262,7 52,207,80,235,0 295,7,337,49,0 34,39,62,67,1 250,110,306,166,2 147,316,189,358,7 139,140,195,196,0 95,14,207,126,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000393.jpg 294,168,316,190,5 30,116,44,130,7 249,212,271,234,2 293,301,335,343,0 195,46,223,74,7 98,87,210,199,0 154,229,266,341,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000394.jpg 162,228,204,270,0 18,82,74,138,5 202,137,244,179,0 98,54,154,110,3 98,223,126,251,5 326,92,382,148,0 320,220,404,304,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000395.jpg 161,176,183,198,3 13,377,35,399,1 384,54,406,76,1 40,129,82,171,2 285,309,313,337,8 99,7,211,119,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000396.jpg 12,13,68,69,3 130,289,186,345,2 165,81,193,109,2 26,137,82,193,3 252,13,280,41,4 177,207,261,291,8 295,51,407,163,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000397.jpg 45,250,87,292,2 195,189,223,217,5 124,186,180,242,6 169,310,253,394,2 221,97,333,209,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000398.jpg 112,183,134,205,1 249,175,277,203,7 184,128,226,170,4 281,69,309,97,0 121,30,205,114,5 129,255,241,367,0 11,31,123,143,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000399.jpg 218,349,240,371,8 380,299,394,313,7 331,208,345,222,2 31,287,59,315,3 33,196,89,252,3 254,245,296,287,3 70,51,126,107,8 95,166,151,222,8 323,129,365,171,3 146,69,258,181,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000400.jpg 374,30,396,52,6 295,22,317,44,9 114,208,156,250,4 161,293,189,321,3 119,308,147,336,5 324,78,352,106,4 245,212,273,240,1 239,264,267,292,3 215,97,327,209,1 78,15,190,127,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000401.jpg 213,220,235,242,4 303,297,317,311,7 153,338,195,380,4 244,50,286,92,9 315,305,371,361,1 23,35,65,77,4 264,161,306,203,2 32,234,144,346,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000402.jpg 291,336,305,350,4 239,217,253,231,0 89,168,145,224,4 351,211,379,239,7 305,177,347,219,0 203,326,259,382,6 6,212,34,240,8 254,5,338,89,1 123,242,207,326,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000403.jpg 149,238,177,266,7 92,202,134,244,0 237,260,321,344,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000404.jpg 340,98,354,112,9 312,50,326,64,7 221,117,243,139,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000405.jpg 294,17,308,31,6 32,249,74,291,2 335,369,363,397,0 67,164,109,206,9 162,153,246,237,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000406.jpg 346,356,368,378,9 207,354,263,410,8 230,41,286,97,9 45,329,101,385,1 380,86,408,114,0 209,149,251,191,9 180,300,236,356,0 145,59,229,143,2 262,217,346,301,7 41,212,153,324,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000407.jpg 153,172,265,284,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000408.jpg 209,239,223,253,0 356,340,398,382,6 125,153,209,237,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000409.jpg 1,126,29,154,6 98,243,140,285,3 275,273,387,385,3 167,53,251,137,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000410.jpg 390,296,404,310,4 214,67,236,89,7 175,291,231,347,4 280,293,336,349,7 7,336,49,378,4 74,91,116,133,7 79,321,107,349,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000411.jpg 213,121,227,135,7 214,238,242,266,3 21,353,77,409,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000412.jpg 71,227,99,255,0 101,52,143,94,2 65,326,93,354,2 88,149,172,233,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000413.jpg 144,233,166,255,2 298,268,382,352,1 53,41,137,125,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000414.jpg 382,35,404,57,2 337,55,365,83,4 12,31,96,115,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000415.jpg 112,179,168,235,9 214,226,326,338,3 66,297,150,381,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000416.jpg 187,129,209,151,0 118,46,146,74,5 192,214,220,242,3 74,242,186,354,8 287,260,399,372,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000417.jpg 150,98,172,120,6 181,364,195,378,8 109,393,131,415,9 252,168,308,224,1 127,334,155,362,6 112,230,140,258,0 338,152,366,180,0 362,345,404,387,9 239,2,351,114,4 156,285,268,397,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000418.jpg 154,111,168,125,7 51,125,107,181,8 222,199,264,241,3 113,161,155,203,0 3,253,45,295,3 354,184,396,226,8 239,84,323,168,3 166,2,250,86,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000419.jpg 241,44,263,66,7 53,257,81,285,4 295,211,379,295,7 91,136,203,248,5 172,260,256,344,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000420.jpg 91,398,105,412,0 295,60,337,102,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000421.jpg 303,272,331,300,6 35,40,119,124,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000422.jpg 256,250,270,264,6 45,111,73,139,2 328,46,412,130,8 280,264,364,348,4 246,100,330,184,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000423.jpg 316,184,330,198,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000424.jpg 58,53,72,67,7 13,103,69,159,2 244,87,272,115,7 268,324,324,380,3 127,298,183,354,3 70,141,182,253,6 325,41,409,125,7 11,260,95,344,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000425.jpg 324,221,346,243,1 261,323,283,345,7 171,41,193,63,7 7,267,35,295,2 81,101,123,143,7 119,261,147,289,1 44,254,100,310,2 329,59,413,143,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000426.jpg 49,297,105,353,4 340,113,396,169,6 4,299,32,327,0 88,145,116,173,8 80,79,122,121,1 34,13,76,55,4 211,88,323,200,2 217,236,329,348,1 110,183,222,295,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000427.jpg 129,212,143,226,6 100,308,142,350,7 291,180,347,236,0 168,217,224,273,4 15,58,43,86,1 122,76,178,132,7 81,217,123,259,9 196,69,308,181,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000428.jpg 321,321,335,335,4 1,9,15,23,6 15,26,71,82,3 146,73,202,129,1 286,99,328,141,0 280,171,336,227,7 214,221,242,249,4 72,14,128,70,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000429.jpg 167,273,189,295,6 294,360,316,382,3 160,369,174,383,0 136,161,178,203,2 365,219,393,247,0 95,49,151,105,5 231,153,273,195,9 286,9,342,65,4 303,150,331,178,4 201,43,285,127,9 18,138,102,222,5 331,302,415,386,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000430.jpg 73,268,115,310,3 290,279,332,321,9 198,172,254,228,7 305,322,333,350,4 379,205,407,233,4 42,241,70,269,4 272,83,356,167,2 298,181,382,265,8 116,36,228,148,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000431.jpg 224,41,252,69,0 298,293,326,321,6 220,221,248,249,3 51,36,163,148,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000432.jpg 78,391,92,405,0 266,100,288,122,6 285,221,299,235,5 300,161,412,273,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000433.jpg 78,235,120,277,8 295,271,337,313,4 375,162,403,190,1 277,80,305,108,8 39,128,95,184,2 346,293,402,349,1 238,149,322,233,6 35,328,119,412,2 126,231,238,343,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000434.jpg 234,75,248,89,8 371,320,399,348,3 209,188,265,244,9 113,101,197,185,3 14,198,98,282,3 91,244,203,356,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000435.jpg 52,271,108,327,6 231,150,259,178,0 258,87,314,143,6 116,176,172,232,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000436.jpg 290,30,318,58,2 192,87,304,199,9 56,234,168,346,2 276,226,360,310,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000437.jpg 54,374,68,388,7 271,43,285,57,7 113,248,169,304,9 115,99,143,127,1 5,231,61,287,6 203,257,245,299,2 192,179,220,207,4 189,7,273,91,5 6,92,118,204,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000438.jpg 229,133,251,155,6 332,294,354,316,3 210,109,224,123,9 63,292,91,320,9 78,211,134,267,0 4,371,32,399,2 96,323,124,351,8 152,32,236,116,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000439.jpg 382,358,396,372,7 65,378,79,392,1 218,175,274,231,0 326,140,368,182,8 9,52,65,108,1 313,241,341,269,0 144,108,228,192,0 6,136,118,248,5 179,15,263,99,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000440.jpg 273,205,329,261,9 70,17,98,45,7 121,253,177,309,2 223,308,279,364,9 96,61,180,145,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000441.jpg 248,6,262,20,4 100,361,122,383,0 371,232,385,246,7 19,318,47,346,9 37,174,93,230,8 117,219,173,275,9 112,100,224,212,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000442.jpg 208,232,222,246,4 263,11,305,53,8 284,188,340,244,3 369,68,397,96,2 44,296,100,352,9 79,246,107,274,2 186,320,270,404,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000443.jpg 283,249,297,263,0 145,311,201,367,2 349,115,405,171,3 365,288,407,330,2 87,65,143,121,9 201,381,229,409,4 247,185,359,297,3 94,170,206,282,1 248,65,332,149,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000444.jpg 361,21,403,63,5 162,55,204,97,9 287,368,315,396,1 63,123,119,179,2 48,237,104,293,9 326,237,410,321,8 270,130,354,214,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000445.jpg 276,236,290,250,7 373,193,387,207,3 273,51,329,107,5 343,75,371,103,3 125,119,237,231,4 48,216,132,300,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000446.jpg 343,163,365,185,5 102,21,124,43,0 57,298,71,312,4 281,165,323,207,8 296,299,338,341,5 357,26,399,68,9 295,55,323,83,3 113,171,225,283,2 186,45,298,157,0 88,302,172,386,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000447.jpg 250,350,292,392,0 177,220,233,276,2 34,337,62,365,3 56,370,84,398,8 258,56,370,168,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000448.jpg 27,276,49,298,6 261,217,275,231,9 177,159,205,187,3 237,4,265,32,4 250,181,278,209,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000449.jpg 128,235,150,257,8 45,359,87,401,2 94,371,136,413,5 129,32,171,74,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000450.jpg 291,387,313,409,5 62,28,90,56,9 139,248,167,276,8 107,197,149,239,7 274,62,316,104,5 183,134,295,246,8 292,129,404,241,2 61,81,173,193,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000451.jpg 160,91,182,113,0 212,347,234,369,9 50,119,134,203,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000452.jpg 218,243,240,265,1 286,255,328,297,2 378,90,406,118,9 151,53,263,165,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000453.jpg 364,372,378,386,7 117,1,131,15,8 129,9,151,31,3 77,267,105,295,5 33,275,75,317,0 307,220,335,248,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000454.jpg 246,321,268,343,3 274,49,288,63,4 192,86,248,142,7 42,343,70,371,5 5,2,33,30,5 141,32,197,88,1 186,228,242,284,0 341,64,369,92,4 323,226,407,310,7 14,193,98,277,1 62,113,146,197,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000455.jpg 280,132,294,146,0 40,378,54,392,6 212,258,324,370,9 34,289,146,401,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000456.jpg 284,315,298,329,6 9,18,31,40,4 174,224,230,280,5 63,306,147,390,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000457.jpg 320,56,334,70,2 179,261,201,283,4 134,294,156,316,0 106,37,162,93,0 256,231,298,273,8 50,292,78,320,3 22,147,78,203,2 94,185,178,269,0 225,286,309,370,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000458.jpg 253,159,309,215,5 368,86,410,128,8 103,56,215,168,6 131,313,215,397,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000459.jpg 361,287,383,309,9 11,361,25,375,5 6,212,28,234,7 246,22,302,78,6 94,279,122,307,3 115,14,143,42,2 292,272,348,328,1 180,286,264,370,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000460.jpg 194,91,216,113,3 311,386,333,408,2 35,77,77,119,1 160,311,244,395,5 213,133,325,245,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000461.jpg 24,247,38,261,9 81,118,103,140,6 328,111,342,125,9 259,368,301,410,4 289,226,331,268,7 363,285,405,327,9 300,289,342,331,7 173,267,229,323,2 91,7,203,119,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000462.jpg 32,401,46,415,5 89,182,173,266,7 276,115,360,199,4 138,46,250,158,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000463.jpg 138,326,222,410,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000464.jpg 76,120,98,142,5 310,0,366,56,4 194,16,222,44,1 144,136,200,192,1 322,226,378,282,6 44,145,72,173,7 182,39,294,151,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000465.jpg 98,27,120,49,3 195,254,209,268,7 63,251,85,273,2 173,76,201,104,5 309,214,337,242,8 203,269,231,297,7 46,159,130,243,0 328,231,412,315,1 322,327,406,411,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000466.jpg 317,378,345,406,7 126,201,182,257,4 224,360,252,388,7 102,267,130,295,8 58,73,142,157,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000467.jpg 375,269,389,283,0 145,143,159,157,8 302,3,358,59,8 248,58,276,86,3 196,155,238,197,2 303,70,331,98,0 81,183,165,267,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000468.jpg 76,244,98,266,5 274,175,330,231,6 98,233,154,289,4 266,4,322,60,6 39,210,81,252,0 279,245,307,273,9 144,373,186,415,8 84,141,168,225,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000469.jpg 179,271,221,313,5 90,340,118,368,7 181,152,237,208,0 268,243,324,299,2 75,139,103,167,8 310,162,394,246,4 266,40,378,152,4 2,212,114,324,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000470.jpg 2,176,24,198,2 365,25,407,67,1 108,242,220,354,6 88,52,200,164,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000471.jpg 188,253,210,275,0 25,247,39,261,9 346,373,368,395,8 329,3,385,59,1 273,114,315,156,4 45,91,73,119,5 108,175,150,217,0 280,172,308,200,5 133,41,217,125,2 21,149,105,233,8 279,214,363,298,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000472.jpg 232,155,274,197,9 330,233,386,289,7 284,91,326,133,8 8,216,50,258,7 151,82,179,110,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000473.jpg 251,8,279,36,4 60,212,102,254,7 5,319,47,361,4 37,10,93,66,3 208,197,236,225,3 69,261,153,345,6 292,124,376,208,9 130,49,242,161,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000474.jpg 81,184,103,206,3 199,326,221,348,2 350,292,372,314,7 97,66,181,150,8 284,170,368,254,1 217,49,329,161,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000475.jpg 103,126,117,140,0 178,223,220,265,4 239,146,281,188,1 152,91,208,147,4 261,2,289,30,9 330,23,414,107,9 273,220,357,304,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000476.jpg 52,359,66,373,5 380,113,402,135,6 123,377,145,399,0 222,237,278,293,6 199,93,255,149,6 94,385,122,413,3 23,291,51,319,0 23,77,65,119,5 295,268,407,380,0 168,293,252,377,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000477.jpg 301,96,323,118,7 253,318,295,360,4 58,288,100,330,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000478.jpg 181,184,237,240,4 59,300,101,342,0 86,151,142,207,9 253,332,295,374,9 130,298,242,410,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000479.jpg 64,115,78,129,9 313,219,335,241,9 78,314,134,370,8 192,91,248,147,5 163,157,191,185,2 188,237,300,349,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000480.jpg 388,2,410,24,1 270,89,292,111,6 78,10,134,66,3 262,30,318,86,7 146,74,202,130,6 25,282,137,394,3 161,289,273,401,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000481.jpg 195,167,209,181,6 383,14,397,28,3 78,349,106,377,1 294,347,336,389,8 238,42,266,70,3 193,318,249,374,3 251,272,307,328,1 60,201,144,285,9 245,101,329,185,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000482.jpg 374,45,396,67,7 79,169,93,183,7 8,14,36,42,5 303,88,359,144,0 195,351,251,407,1 97,248,209,360,8 48,29,160,141,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000483.jpg 379,223,401,245,5 114,176,142,204,0 205,211,233,239,1 286,15,342,71,7 130,50,158,78,0 164,240,248,324,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000484.jpg 115,198,157,240,9 5,279,47,321,2 235,244,347,356,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000485.jpg 136,371,158,393,2 274,375,296,397,9 107,267,149,309,1 14,244,70,300,1 55,75,111,131,2 236,151,278,193,6 330,189,386,245,5 102,28,158,84,1 132,85,216,169,8 303,244,415,356,1 244,40,356,152,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000486.jpg 194,186,208,200,5 231,179,259,207,8 77,73,119,115,5 58,277,100,319,9 227,292,283,348,9 84,167,168,251,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000487.jpg 328,321,342,335,3 386,370,408,392,8 204,67,316,179,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000488.jpg 315,202,337,224,7 9,279,37,307,3 192,320,234,362,1 337,64,379,106,9 234,258,346,370,7 35,303,147,415,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000489.jpg 236,17,258,39,3 384,370,406,392,5 226,109,248,131,3 355,260,397,302,5 91,308,119,336,5 58,134,114,190,3 357,47,399,89,6 1,190,43,232,0 364,155,406,197,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000490.jpg 343,81,365,103,8 112,315,168,371,2 13,96,41,124,2 275,130,303,158,7 239,62,267,90,3 93,50,121,78,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000491.jpg 364,237,378,251,6 213,43,227,57,9 203,244,225,266,1 34,309,90,365,1 341,116,369,144,6 232,18,274,60,8 327,233,355,261,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000492.jpg 294,67,316,89,1 25,48,81,104,0 258,320,300,362,7 320,215,404,299,6 318,130,402,214,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000493.jpg 241,199,263,221,4 33,65,89,121,4 165,213,221,269,7 101,270,129,298,2 190,268,246,324,9 273,69,357,153,5 41,301,153,413,9 240,296,352,408,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000494.jpg 188,333,210,355,6 138,130,166,158,3 285,176,341,232,3 255,114,283,142,0 22,88,64,130,1 228,280,284,336,6 87,239,143,295,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000495.jpg 43,335,57,349,7 370,28,384,42,8 215,367,229,381,2 307,33,391,117,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000496.jpg 91,255,147,311,9 284,73,340,129,8 245,30,273,58,0 119,341,175,397,0 287,184,343,240,7 211,139,253,181,7 265,237,377,349,7 41,100,153,212,5 154,203,266,315,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000497.jpg 292,259,306,273,4 358,241,372,255,6 276,281,290,295,1 297,386,325,414,7 14,212,56,254,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000498.jpg 320,337,334,351,5 152,91,166,105,4 42,273,70,301,4 291,377,319,405,1 83,44,125,86,7 331,255,387,311,1 109,132,165,188,8 132,332,188,388,5 277,124,389,236,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000499.jpg 391,134,405,148,8 184,107,198,121,6 91,326,147,382,6 142,230,184,272,4 196,18,238,60,1 56,141,84,169,9 310,72,394,156,7 59,8,171,120,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000500.jpg 149,228,177,256,8 12,311,54,353,3 353,162,395,204,4 313,116,341,144,0 123,88,165,130,7 84,250,196,362,6 249,31,333,115,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000501.jpg 212,129,240,157,6 225,78,253,106,1 1,107,29,135,5 50,65,134,149,7 109,155,193,239,3 26,251,110,335,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000502.jpg 258,320,280,342,2 142,252,156,266,6 88,374,110,396,5 236,238,292,294,2 64,127,120,183,6 276,194,318,236,9 194,36,306,148,4 307,275,391,359,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000503.jpg 363,373,391,401,8 25,217,67,259,1 272,84,300,112,4 157,7,241,91,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000504.jpg 194,130,278,214,1 275,5,359,89,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000505.jpg 102,330,116,344,9 125,360,139,374,6 122,241,178,297,9 185,70,213,98,0 154,108,210,164,4 253,237,365,349,2 174,201,258,285,5 236,91,348,203,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000506.jpg 199,247,221,269,4 186,186,208,208,2 163,238,185,260,5 259,297,287,325,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000507.jpg 184,243,240,299,3 33,108,61,136,0 23,53,65,95,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000508.jpg 241,319,263,341,2 293,185,315,207,4 364,69,378,83,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000509.jpg 314,233,328,247,8 32,252,46,266,9 79,180,101,202,8 120,146,204,230,7 194,26,278,110,9 118,268,202,352,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000510.jpg 28,359,42,373,6 299,358,355,414,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000511.jpg 168,391,182,405,2 278,112,300,134,1 152,332,208,388,1 157,222,213,278,5 220,170,262,212,3 311,170,353,212,7 3,243,31,271,4 333,40,375,82,9 75,72,187,184,1 209,320,293,404,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000512.jpg 355,236,369,250,7 180,191,202,213,4 185,255,207,277,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000513.jpg 210,151,232,173,1 109,203,131,225,1 290,65,332,107,3 6,221,48,263,0 125,284,153,312,5 179,174,221,216,5 3,270,59,326,1 23,324,107,408,4 215,202,327,314,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000514.jpg 109,163,131,185,9 19,301,33,315,1 256,329,270,343,5 307,278,349,320,7 194,184,222,212,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000515.jpg 314,87,336,109,6 347,290,361,304,4 18,24,74,80,1 257,236,299,278,5 260,60,288,88,3 199,100,255,156,8 324,242,366,284,0 117,346,159,388,1 37,231,149,343,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000516.jpg 215,99,229,113,5 35,146,91,202,7 129,92,185,148,7 52,257,136,341,3 188,88,300,200,1 243,209,355,321,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000517.jpg 291,320,313,342,4 1,18,15,32,1 171,306,185,320,2 99,262,127,290,0 142,207,184,249,3 149,253,191,295,2 370,66,412,108,1 186,73,270,157,0 75,300,187,412,5 204,328,288,412,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000518.jpg 128,2,142,16,2 306,191,320,205,7 4,5,116,117,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000519.jpg 198,197,254,253,6 51,181,107,237,0 316,148,358,190,3 62,155,90,183,0 278,305,334,361,1 155,196,197,238,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000520.jpg 211,29,225,43,1 343,53,371,81,9 188,238,230,280,1 53,158,165,270,2 136,53,220,137,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000521.jpg 316,62,372,118,1 243,27,271,55,5 204,316,246,358,0 357,26,385,54,6 85,171,169,255,3 245,208,329,292,5 65,20,177,132,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000522.jpg 48,173,70,195,4 277,279,305,307,0 174,291,230,347,2 355,271,411,327,7 294,378,322,406,7 328,152,384,208,8 118,187,146,215,0 59,79,143,163,2 35,292,147,404,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000523.jpg 181,21,223,63,9 130,265,158,293,2 198,65,240,107,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000524.jpg 213,183,255,225,6 311,288,367,344,6 228,105,256,133,1 335,172,363,200,7 59,183,115,239,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000525.jpg 110,311,124,325,5 284,265,298,279,1 42,193,56,207,6 316,94,358,136,2 259,339,287,367,9 363,130,391,158,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000526.jpg 239,159,253,173,8 295,323,309,337,7 227,328,269,370,5 333,88,389,144,3 66,22,108,64,5 283,28,339,84,8 109,154,151,196,3 327,219,411,303,0 172,152,284,264,3 99,56,183,140,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000527.jpg 60,78,74,92,0 330,265,372,307,0 159,63,187,91,4 228,53,256,81,4 169,297,197,325,4 344,183,372,211,1 308,128,364,184,9 193,311,277,395,1 216,89,300,173,1 86,191,198,303,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000528.jpg 305,380,319,394,9 227,159,241,173,4 381,83,403,105,9 266,345,294,373,9 318,162,374,218,5 42,15,98,71,7 272,383,300,411,7 281,239,365,323,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000529.jpg 108,128,122,142,2 54,236,76,258,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000530.jpg 128,268,150,290,8 203,330,225,352,3 152,58,166,72,1 247,37,289,79,0 42,47,70,75,2 36,228,120,312,8 231,183,315,267,5 253,73,365,185,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000531.jpg 27,144,49,166,2 98,167,154,223,7 358,339,386,367,3 202,96,230,124,0 339,30,395,86,7 30,366,58,394,2 165,246,193,274,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000532.jpg 334,335,356,357,2 124,142,180,198,5 277,97,333,153,3 166,312,194,340,8 202,51,244,93,9 57,61,141,145,9 189,276,301,388,3 12,202,124,314,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000533.jpg 148,137,162,151,0 293,380,321,408,7 68,105,124,161,9 121,182,177,238,7 75,64,117,106,1 120,332,148,360,0 341,20,369,48,3 285,218,397,330,3 209,4,293,88,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000534.jpg 281,262,295,276,1 229,136,285,192,7 169,95,225,151,7 358,385,386,413,1 383,184,411,212,6 10,268,66,324,8 73,192,115,234,0 25,93,109,177,4 317,70,401,154,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000535.jpg 45,310,73,338,1 329,131,357,159,9 176,265,204,293,8 335,313,391,369,0 96,337,138,379,1 97,126,181,210,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000536.jpg 210,265,224,279,6 209,60,237,88,6 132,78,188,134,5 45,277,87,319,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000537.jpg 376,379,398,401,2 43,376,65,398,1 232,36,260,64,5 291,14,347,70,0 320,330,348,358,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000538.jpg 121,181,143,203,0 15,154,29,168,5 167,212,181,226,0 343,222,399,278,6 200,340,242,382,9 163,83,205,125,3 34,183,90,239,6 297,2,339,44,2 308,76,364,132,1 201,115,313,227,3 50,80,134,164,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000539.jpg 308,117,330,139,5 70,350,84,364,6 107,302,121,316,1 337,136,393,192,9 279,30,363,114,9 113,144,197,228,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000540.jpg 355,325,377,347,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000541.jpg 290,312,312,334,5 239,268,261,290,6 284,55,312,83,5 92,138,148,194,0 146,53,188,95,9 352,283,394,325,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000542.jpg 223,320,265,362,3 172,64,214,106,3 310,74,366,130,2 99,151,155,207,6 26,92,82,148,3 130,358,186,414,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000543.jpg 72,57,86,71,2 64,215,106,257,4 11,367,39,395,9 317,258,345,286,4 340,341,396,397,3 201,13,313,125,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000544.jpg 270,306,326,362,5 45,89,73,117,9 158,186,214,242,9 160,94,202,136,1 167,35,223,91,4 314,158,370,214,5 202,110,286,194,8 133,244,245,356,7 4,232,116,344,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000545.jpg 147,264,161,278,4 251,309,265,323,1 367,326,389,348,2 81,329,137,385,9 175,88,203,116,7 181,219,293,331,4 37,7,149,119,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000546.jpg 298,196,312,210,5 388,83,410,105,7 93,115,149,171,0 336,187,378,229,2 188,210,216,238,0 161,283,273,395,1 282,223,394,335,6 51,7,163,119,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000547.jpg 316,260,338,282,0 159,254,181,276,7 264,289,286,311,8 134,52,162,80,0 4,193,46,235,5 138,152,180,194,8 262,352,290,380,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000548.jpg 255,91,297,133,8 134,147,190,203,3 261,4,289,32,9 145,231,257,343,2 303,20,387,104,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000549.jpg 114,23,136,45,5 14,58,36,80,2 387,277,401,291,6 82,120,138,176,1 45,226,157,338,3 310,27,394,111,3 271,230,383,342,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000550.jpg 363,188,377,202,7 216,287,244,315,9 159,128,215,184,5 51,39,93,81,7 74,207,158,291,7 62,105,146,189,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000551.jpg 191,5,213,27,1 72,346,94,368,4 106,286,162,342,7 360,115,388,143,1 350,8,378,36,5 189,368,217,396,4 314,253,398,337,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000552.jpg 307,123,391,207,1 78,176,190,288,0 183,45,267,129,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000553.jpg 182,385,196,399,0 184,249,198,263,8 306,315,328,337,8 323,358,379,414,3 323,42,365,84,6 199,75,227,103,1 92,141,134,183,3 160,15,216,71,1 131,273,187,329,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000554.jpg 111,173,125,187,6 25,247,39,261,6 231,12,259,40,1 203,270,245,312,9 229,44,341,156,9 37,229,149,341,8 81,38,165,122,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000555.jpg 197,289,219,311,1 195,337,237,379,0 125,17,167,59,1 26,217,82,273,5 39,101,95,157,4 80,316,108,344,3 91,286,119,314,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000556.jpg 270,285,312,327,4 310,81,352,123,1 193,263,235,305,3 304,179,360,235,7 11,275,123,387,8 117,143,201,227,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000557.jpg 61,132,117,188,1 183,279,239,335,0 322,98,378,154,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000558.jpg 25,366,39,380,9 234,132,290,188,3 111,265,167,321,6 177,292,289,404,9 302,87,414,199,9 135,21,247,133,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000559.jpg 288,190,310,212,9 382,262,404,284,4 97,336,111,350,2 227,273,339,385,4 162,68,246,152,4 104,210,188,294,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000560.jpg 192,239,234,281,2 292,17,334,59,8 67,138,109,180,7 161,321,203,363,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000561.jpg 148,352,170,374,8 280,16,336,72,6 178,134,220,176,4 20,144,76,200,7 217,46,245,74,5 91,188,147,244,2 220,89,262,131,6 244,135,356,247,0 184,295,296,407,3 76,80,160,164,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000562.jpg 283,138,305,160,6 24,186,46,208,5 140,202,154,216,9 133,250,161,278,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000563.jpg 328,377,350,399,8 264,179,286,201,7 73,215,95,237,8 237,275,265,303,0 271,0,313,42,2 175,264,203,292,9 23,354,51,382,1 325,180,353,208,8 1,246,57,302,8 125,91,209,175,4 65,245,177,357,1 292,52,404,164,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000564.jpg 12,228,26,242,3 250,141,306,197,7 343,136,385,178,8 222,107,250,135,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000565.jpg 81,379,95,393,6 300,369,322,391,3 215,58,243,86,3 346,168,402,224,0 286,303,314,331,8 210,157,238,185,5 225,235,253,263,1 203,268,231,296,6 22,118,134,230,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000566.jpg 158,33,270,145,5 159,283,243,367,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000567.jpg 291,29,403,141,4 74,252,158,336,0 193,266,305,378,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000568.jpg 83,155,195,267,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000569.jpg 43,73,65,95,9 305,365,347,407,7 13,108,69,164,1 369,338,411,380,8 190,373,218,401,6 112,61,168,117,8 76,133,160,217,1 289,274,373,358,3 269,82,381,194,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000570.jpg 359,385,373,399,2 159,195,215,251,0 301,271,329,299,4 84,174,140,230,9 386,208,414,236,3 44,68,128,152,5 196,246,308,358,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000571.jpg 337,77,351,91,5 105,300,127,322,2 306,13,320,27,7 62,162,118,218,1 136,210,164,238,1 276,162,388,274,9 130,67,214,151,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000572.jpg 73,235,87,249,6 82,389,104,411,1 10,181,32,203,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000573.jpg 286,79,300,93,8 141,16,155,30,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000574.jpg 352,243,374,265,2 129,202,151,224,7 30,234,44,248,7 108,272,164,328,7 309,279,337,307,3 60,12,102,54,1 53,97,81,125,3 200,3,242,45,9 323,77,365,119,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000575.jpg 168,351,182,365,6 107,256,121,270,9 279,166,321,208,7 206,6,318,118,7 59,72,143,156,0 150,246,234,330,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000576.jpg 388,377,410,399,1 203,233,225,255,7 250,164,306,220,7 226,263,254,291,3 269,266,325,322,5 105,145,133,173,7 134,24,246,136,8 17,211,129,323,1 3,41,115,153,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000577.jpg 144,55,186,97,6 180,312,208,340,1 214,310,270,366,1 251,112,279,140,7 112,132,168,188,7 229,3,313,87,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000578.jpg 295,89,351,145,7 99,295,127,323,6 146,124,188,166,9 115,60,171,116,8 343,27,385,69,8 188,101,230,143,1 331,315,415,399,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000579.jpg 94,383,116,405,6 123,376,137,390,0 48,366,70,388,6 219,65,303,149,2 115,146,227,258,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000580.jpg 290,190,312,212,8 378,348,406,376,0 85,196,141,252,3 168,230,224,286,9 197,302,239,344,1 289,99,331,141,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000581.jpg 8,329,36,357,5 242,33,298,89,1 16,359,58,401,8 250,241,306,297,6 44,249,72,277,6 247,115,331,199,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000582.jpg 271,105,293,127,0 105,96,161,152,8 201,259,257,315,4 188,369,216,397,3 25,162,109,246,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000583.jpg 252,149,294,191,4 292,32,348,88,8 110,55,222,167,6 92,287,176,371,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000584.jpg 181,152,195,166,1 207,326,221,340,3 343,331,399,387,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000585.jpg 169,202,191,224,8 36,240,58,262,6 87,230,129,272,2 290,315,346,371,6 353,366,395,408,6 58,188,100,230,8 277,210,361,294,9 144,252,228,336,3 255,30,367,142,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000586.jpg 242,351,284,393,9 128,174,240,286,4 239,220,351,332,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000587.jpg 134,18,148,32,1 330,46,352,68,1 33,266,61,294,4 239,272,351,384,6 128,275,240,387,2 106,177,190,261,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000588.jpg 148,215,170,237,8 375,40,397,62,4 195,350,251,406,9 60,0,116,56,1 70,124,98,152,7 334,100,390,156,2 145,102,257,214,4 13,191,97,275,2 113,261,197,345,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000589.jpg 30,248,44,262,7 286,285,308,307,9 191,196,275,280,5 285,142,369,226,2 113,18,197,102,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000590.jpg 285,88,313,116,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000591.jpg 190,205,204,219,5 106,180,128,202,8 118,185,230,297,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000592.jpg 167,145,181,159,4 176,62,198,84,5 350,64,364,78,7 370,220,398,248,3 8,286,92,370,7 85,219,197,331,4 89,85,173,169,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000593.jpg 244,113,300,169,3 61,295,89,323,7 140,43,224,127,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000594.jpg 9,323,23,337,4 174,164,196,186,4 79,305,93,319,1 208,328,264,384,1 154,262,196,304,3 337,164,365,192,7 19,245,61,287,9 119,219,147,247,5 220,169,332,281,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000595.jpg 11,148,33,170,7 391,392,413,414,1 153,374,181,402,1 192,345,234,387,4 94,174,178,258,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000596.jpg 173,74,187,88,9 101,40,143,82,7 222,274,306,358,7 56,212,140,296,9 268,0,380,112,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000597.jpg 142,286,164,308,4 16,252,128,364,1 34,63,146,175,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000598.jpg 386,253,408,275,6 358,343,380,365,7 39,68,53,82,7 102,86,158,142,3 74,356,130,412,9 220,38,262,80,5 49,152,91,194,0 343,118,371,146,0 152,18,180,46,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000599.jpg 306,184,362,240,9 106,299,134,327,6 3,175,115,287,7 214,157,298,241,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000600.jpg 331,184,359,212,0 187,231,243,287,0 129,109,185,165,9 363,99,405,141,3 81,128,123,170,7 33,285,117,369,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000601.jpg 111,271,133,293,4 263,100,285,122,6 162,192,218,248,5 217,291,301,375,2 220,19,304,103,2 226,131,338,243,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000602.jpg 35,163,57,185,4 57,91,79,113,7 49,268,63,282,6 176,118,218,160,7 181,64,237,120,3 287,338,315,366,4 269,114,297,142,0 129,275,157,303,8 338,224,380,266,7 230,171,314,255,0 99,167,211,279,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000603.jpg 118,212,132,226,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000604.jpg 343,159,371,187,3 293,200,321,228,8 106,102,162,158,8 38,91,80,133,8 155,66,197,108,2 28,308,112,392,5 86,195,198,307,2 271,250,383,362,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000605.jpg 174,365,216,407,8 200,271,242,313,7 96,314,152,370,3 192,115,234,157,4 286,99,398,211,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000606.jpg 344,35,366,57,0 295,98,337,140,3 180,21,264,105,4 320,160,404,244,7 235,238,347,350,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000607.jpg 361,56,383,78,8 11,228,39,256,7 75,332,131,388,3 173,285,201,313,3 91,259,147,315,4 328,299,412,383,4 45,82,157,194,0 189,15,273,99,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000608.jpg 227,33,283,89,4 98,212,126,240,1 12,182,54,224,6 285,183,397,295,7 41,38,153,150,1 205,199,289,283,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000609.jpg 309,220,331,242,6 185,353,241,409,7 29,147,85,203,8 327,303,383,359,1 107,221,135,249,3 381,205,409,233,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000610.jpg 35,363,49,377,1 180,377,208,405,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000611.jpg 96,168,110,182,9 393,204,415,226,4 83,306,97,320,8 129,70,157,98,9 183,14,225,56,0 163,109,191,137,6 227,6,283,62,3 99,277,211,389,1 60,145,172,257,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000612.jpg 158,196,186,224,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000613.jpg 151,247,207,303,8 311,209,367,265,4 84,231,140,287,5 319,25,347,53,4 246,260,302,316,5 297,329,353,385,8 4,68,116,180,0 236,63,320,147,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000614.jpg 262,136,276,150,3 303,285,345,327,2 220,257,248,285,7 216,54,272,110,9 280,142,392,254,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000615.jpg 206,173,220,187,7 76,289,90,303,3 15,166,71,222,1 353,138,395,180,5 282,37,310,65,1 345,316,373,344,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000616.jpg 332,74,354,96,3 330,168,344,182,2 172,316,200,344,2 52,343,94,385,7 330,284,358,312,7 349,149,391,191,2 14,182,98,266,1 154,171,238,255,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000617.jpg 375,245,397,267,2 11,339,67,395,6 101,304,129,332,8 197,159,309,271,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000618.jpg 357,102,399,144,7 293,330,377,414,5 127,125,239,237,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000619.jpg 190,273,204,287,1 239,263,281,305,3 103,200,159,256,7 240,84,324,168,1 70,70,154,154,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000620.jpg 375,178,397,200,9 248,279,262,293,3 126,145,148,167,0 166,172,222,228,1 318,238,374,294,0 42,10,98,66,4 98,32,140,74,9 321,95,363,137,9 294,292,406,404,8 44,196,156,308,8 55,314,139,398,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000621.jpg 103,239,131,267,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000622.jpg 331,313,359,341,6 6,247,62,303,2 331,5,359,33,4 213,264,269,320,1 56,323,98,365,7 228,206,256,234,1 251,89,335,173,1 38,94,122,178,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000623.jpg 388,275,402,289,9 168,283,190,305,6 280,156,302,178,4 369,116,397,144,5 86,328,128,370,9 65,92,121,148,8 226,115,268,157,2 293,256,377,340,7 19,181,131,293,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000624.jpg 302,374,324,396,1 107,322,121,336,1 182,83,196,97,6 284,241,312,269,3 18,275,46,303,7 86,140,198,252,4 183,275,295,387,7 68,24,152,108,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000625.jpg 337,357,359,379,9 163,332,219,388,6 371,62,399,90,4 88,180,116,208,3 54,22,138,106,7 329,247,413,331,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000626.jpg 10,154,32,176,2 40,60,54,74,2 344,244,358,258,4 287,329,329,371,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000627.jpg 301,288,323,310,0 353,155,409,211,0 100,203,156,259,1 130,285,172,327,0 308,84,336,112,9 248,303,304,359,2 180,208,236,264,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000628.jpg 251,60,265,74,2 227,282,249,304,8 228,115,242,129,4 247,220,331,304,2 34,128,146,240,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000629.jpg 127,34,141,48,3 130,287,172,329,5 37,83,149,195,3 208,182,320,294,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000630.jpg 14,276,28,290,9 371,145,385,159,3 342,146,356,160,1 210,98,266,154,7 318,48,374,104,1 264,222,320,278,7 231,38,273,80,0 43,267,85,309,9 36,312,120,396,4 112,114,196,198,9 308,308,392,392,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000631.jpg 92,88,120,116,1 175,281,217,323,0 298,101,354,157,7 190,176,232,218,4 47,320,131,404,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000632.jpg 236,276,292,332,9 125,212,167,254,8 227,68,269,110,3 108,51,220,163,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000633.jpg 102,22,124,44,7 306,20,328,42,1 206,138,248,180,0 169,151,197,179,6 150,281,192,323,3 246,190,288,232,6 0,323,56,379,7 102,39,214,151,9 35,174,147,286,7 227,246,339,358,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000634.jpg 206,254,290,338,1 46,62,158,174,5 118,290,202,374,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000635.jpg 235,86,257,108,9 19,144,33,158,6 49,355,63,369,9 154,260,182,288,0 297,58,339,100,1 298,330,382,414,6 19,268,131,380,4 209,167,293,251,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000636.jpg 218,122,246,150,3 25,2,53,30,8 278,270,334,326,6 214,27,256,69,6 209,332,251,374,0 29,284,71,326,2 242,54,354,166,0 120,323,204,407,8 108,11,220,123,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000637.jpg 355,278,377,300,2 98,36,126,64,1 319,93,375,149,7 0,59,42,101,8 49,179,91,221,7 25,247,53,275,4 156,109,240,193,8 17,270,101,354,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000638.jpg 106,241,134,269,8 87,15,129,57,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000639.jpg 261,327,283,349,6 388,136,402,150,8 224,178,238,192,1 156,327,184,355,7 332,41,388,97,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000640.jpg 51,82,73,104,7 230,28,244,42,2 290,333,304,347,7 329,312,371,354,5 246,256,330,340,3 243,73,355,185,8 33,206,117,290,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000641.jpg 245,291,287,333,7 45,66,101,122,9 92,201,148,257,1 301,83,357,139,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000642.jpg 202,6,216,20,3 11,70,33,92,1 25,182,39,196,7 85,80,113,108,8 161,140,245,224,4 100,295,212,407,5 245,150,357,262,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000643.jpg 235,142,257,164,7 112,290,168,346,3 0,191,84,275,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000644.jpg 190,229,204,243,8 246,91,268,113,8 312,179,354,221,9 337,220,379,262,3 113,337,141,365,5 267,347,309,389,9 100,49,142,91,7 40,239,82,281,3 182,236,294,348,4 286,49,370,133,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000645.jpg 277,333,299,355,1 333,378,361,406,0 142,173,184,215,5 290,76,374,160,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000646.jpg 99,246,121,268,5 335,192,363,220,8 190,350,218,378,9 257,77,285,105,5 318,218,360,260,9 339,358,381,400,9 25,79,137,191,5 192,1,276,85,1 179,192,291,304,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000647.jpg 340,364,362,386,6 165,16,207,58,4 290,201,332,243,2 39,45,95,101,5 235,94,319,178,5 135,262,247,374,2 103,109,187,193,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000648.jpg 9,112,23,126,2 52,346,66,360,9 84,366,98,380,3 28,8,112,92,8 82,201,194,313,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000649.jpg 343,135,357,149,9 381,266,403,288,8 307,52,363,108,5 157,39,185,67,3 201,41,243,83,6 231,259,287,315,0 80,271,122,313,4 13,33,97,117,3 97,113,181,197,4 201,141,313,253,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000650.jpg 220,345,234,359,7 177,377,199,399,1 127,300,141,314,2 155,58,197,100,1 43,97,99,153,4 198,65,226,93,0 229,136,313,220,0 49,284,161,396,0 113,172,225,284,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000651.jpg 89,326,103,340,2 284,8,298,22,4 74,199,88,213,2 267,55,309,97,2 37,7,65,35,4 302,355,344,397,1 153,95,209,151,0 275,202,359,286,8 135,227,247,339,1 24,137,136,249,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000652.jpg 113,128,169,184,0 236,323,292,379,1 213,46,325,158,8 102,3,214,115,0 254,164,366,276,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000653.jpg 295,353,309,367,2 162,329,218,385,6 68,297,152,381,6 309,259,393,343,8 125,170,237,282,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000654.jpg 245,278,267,300,4 264,344,286,366,5 300,195,322,217,7 275,112,331,168,4 253,302,295,344,2 167,209,209,251,3 26,355,68,397,2 25,186,137,298,2 83,13,167,97,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000655.jpg 81,181,109,209,2 286,84,328,126,4 28,195,70,237,1 101,75,213,187,4 241,133,325,217,6 241,276,353,388,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000656.jpg 304,131,318,145,6 369,282,391,304,5 55,361,69,375,4 124,151,166,193,8 276,44,360,128,5 252,187,336,271,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000657.jpg 368,46,396,74,3 111,79,167,135,3 205,103,233,131,3 132,296,160,324,2 328,210,356,238,9 187,172,215,200,3 228,267,340,379,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000658.jpg 163,215,177,229,4 119,8,133,22,1 143,302,165,324,3 327,26,383,82,5 338,96,380,138,2 4,335,46,377,7 318,299,360,341,6 262,297,318,353,8 296,190,380,274,2 162,106,246,190,9 35,127,147,239,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000659.jpg 28,42,42,56,2 117,180,139,202,2 309,97,331,119,4 234,294,276,336,4 78,293,120,335,0 142,348,170,376,5 96,111,152,167,3 195,145,307,257,7 329,175,413,259,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000660.jpg 77,351,99,373,6 287,147,301,161,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000661.jpg 183,289,197,303,2 194,256,216,278,6 116,296,158,338,6 210,354,252,396,2 321,284,377,340,2 103,345,159,401,5 80,155,192,267,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000662.jpg 155,125,177,147,6 267,170,289,192,3 102,270,144,312,1 374,26,402,54,0 306,237,334,265,0 188,231,230,273,1 129,342,157,370,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000663.jpg 211,95,239,123,1 67,46,95,74,1 134,152,162,180,0 330,180,372,222,5 104,267,188,351,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000664.jpg 296,77,338,119,6 207,14,249,56,5 25,114,67,156,6 47,269,103,325,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000665.jpg 142,282,164,304,9 21,56,35,70,6 114,142,198,226,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000666.jpg 73,343,101,371,3 39,31,151,143,3 233,23,317,107,9 10,144,122,256,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000667.jpg 320,354,334,368,1 173,58,195,80,3 193,167,235,209,2 219,273,261,315,5 372,187,414,229,2 48,294,90,336,1 330,223,358,251,3 295,318,323,346,3 86,79,170,163,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000668.jpg 360,160,388,188,0 168,29,210,71,4 135,226,219,310,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000669.jpg 304,88,332,116,6 54,258,82,286,3 255,146,339,230,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000670.jpg 121,21,143,43,8 118,381,132,395,5 240,193,254,207,0 352,198,394,240,5 169,202,225,258,7 231,265,273,307,5 2,345,58,401,9 141,347,197,403,7 161,0,273,112,1 90,114,174,198,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000671.jpg 64,359,78,373,3 211,231,233,253,7 245,251,259,265,8 151,126,207,182,5 120,8,232,120,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000672.jpg 62,26,84,48,6 325,142,339,156,0 298,220,354,276,7 352,305,394,347,5 359,107,387,135,1 316,13,372,69,6 136,172,192,228,2 157,249,269,361,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000673.jpg 313,258,369,314,0 200,94,256,150,1 127,344,155,372,1 31,122,87,178,1 106,114,190,198,8 222,156,306,240,7 188,11,272,95,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000674.jpg 44,370,66,392,1 76,240,90,254,7 230,340,286,396,9 96,230,180,314,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000675.jpg 73,270,87,284,5 167,390,181,404,3 80,63,94,77,2 110,256,138,284,9 317,150,359,192,6 139,111,167,139,4 85,123,127,165,9 15,14,57,56,7 290,351,332,393,2 205,133,289,217,9 181,309,265,393,1 7,168,119,280,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000676.jpg 49,355,63,369,9 175,14,203,42,7 78,90,106,118,1 313,167,397,251,2 58,320,142,404,7 176,164,288,276,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000677.jpg 153,345,195,387,6 328,110,370,152,9 111,32,167,88,1 72,158,128,214,6 282,285,338,341,6 182,97,294,209,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000678.jpg 154,48,168,62,5 20,356,34,370,6 137,149,159,171,2 179,104,291,216,8 155,259,267,371,8 38,280,150,392,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000679.jpg 221,255,235,269,1 193,364,207,378,3 269,344,283,358,1 291,169,333,211,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000680.jpg 70,177,126,233,1 237,336,293,392,0 40,41,82,83,9 324,310,380,366,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000681.jpg 279,354,301,376,6 134,189,176,231,4 19,328,75,384,4 285,15,327,57,6 203,153,287,237,9 287,71,399,183,1 48,201,132,285,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000682.jpg 134,115,190,171,9 295,100,323,128,7 358,329,386,357,9 230,296,272,338,0 293,328,321,356,3 20,22,62,64,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000683.jpg 345,206,359,220,8 377,2,391,16,6 222,136,236,150,5 24,300,66,342,4 266,329,322,385,6 131,147,187,203,0 148,312,176,340,6 95,322,123,350,8 21,184,63,226,6 280,79,392,191,1 118,9,230,121,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000684.jpg 142,51,156,65,1 233,220,247,234,5 354,254,410,310,8 95,300,137,342,4 71,161,99,189,1 0,4,28,32,1 235,19,291,75,9 197,192,309,304,2 115,77,227,189,1 235,78,347,190,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000685.jpg 41,107,83,149,4 90,42,132,84,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000686.jpg 209,11,231,33,7 9,21,51,63,2 187,263,229,305,4 330,150,372,192,1 0,213,28,241,3 38,222,66,250,2 168,155,280,267,2 54,229,166,341,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000687.jpg 137,6,159,28,3 17,253,31,267,3 297,152,339,194,1 5,145,47,187,2 219,275,275,331,9 380,203,408,231,9 290,206,374,290,4 90,327,174,411,3 197,140,281,224,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000688.jpg 371,383,385,397,5 136,49,150,63,9 134,86,162,114,3 29,176,113,260,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000689.jpg 207,52,229,74,1 190,85,246,141,9 164,338,220,394,5 85,349,141,405,1 115,306,157,348,6 248,109,360,221,1 18,16,102,100,8 109,180,221,292,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000690.jpg 305,27,327,49,2 209,292,223,306,3 178,58,192,72,8 63,125,91,153,1 164,178,220,234,1 358,254,414,310,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000691.jpg 163,43,191,71,8 146,232,188,274,7 272,176,314,218,5 33,118,89,174,8 33,36,75,78,7 43,252,155,364,4 211,34,323,146,3 279,237,363,321,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000692.jpg 299,180,313,194,0 35,310,49,324,9 240,312,296,368,0 94,153,136,195,9 25,328,109,412,6 147,138,231,222,1 288,80,372,164,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000693.jpg 241,107,255,121,5 309,238,331,260,9 103,93,159,149,8 206,295,262,351,8 293,90,321,118,1 96,187,180,271,8 229,209,313,293,3 169,47,281,159,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000694.jpg 62,369,84,391,2 385,249,413,277,3 182,295,266,379,9 327,171,411,255,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000695.jpg 95,256,137,298,6 197,86,239,128,6 292,317,334,359,9 187,174,243,230,5 275,244,331,300,5 149,300,261,412,3 252,106,364,218,7 85,24,197,136,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000696.jpg 127,322,149,344,1 355,120,383,148,1 2,84,30,112,3 128,110,156,138,1 273,207,329,263,4 343,152,385,194,1 212,266,268,322,1 6,253,90,337,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000697.jpg 81,268,137,324,8 317,105,373,161,6 215,183,271,239,3 307,158,391,242,3 84,97,196,209,3 147,268,259,380,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000698.jpg 78,266,92,280,4 219,298,233,312,4 131,9,173,51,1 233,94,275,136,3 34,2,76,44,8 149,94,205,150,2 336,119,378,161,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000699.jpg 259,235,281,257,1 114,127,156,169,2 319,298,361,340,7 14,197,70,253,9 316,61,400,145,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000700.jpg 301,303,323,325,1 73,381,87,395,4 371,283,393,305,2 192,218,248,274,9 42,157,84,199,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000701.jpg 371,184,393,206,8 153,249,265,361,2 53,65,165,177,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000702.jpg 16,284,30,298,5 256,182,298,224,7 251,63,279,91,4 239,279,295,335,1 177,265,219,307,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000703.jpg 16,264,72,320,9 15,318,99,402,7 198,5,282,89,9 286,298,398,410,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000704.jpg 289,121,317,149,3 264,332,306,374,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000705.jpg 2,335,24,357,2 151,245,165,259,8 303,212,317,226,5 132,59,188,115,7 156,20,198,62,0 71,223,113,265,1 242,326,298,382,8 33,322,75,364,2 167,132,251,216,8 255,142,367,254,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000706.jpg 399,17,413,31,2 263,128,285,150,5 32,22,116,106,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000707.jpg 206,170,228,192,9 180,213,194,227,0 260,78,282,100,3 76,42,188,154,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000708.jpg 75,246,97,268,7 143,321,227,405,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000709.jpg 269,342,283,356,1 354,344,382,372,8 205,277,261,333,7 292,258,334,300,4 106,264,162,320,0 89,208,145,264,0 347,150,403,206,3 263,10,375,122,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000710.jpg 71,148,93,170,5 2,67,24,89,9 3,333,25,355,4 189,170,217,198,5 20,15,76,71,8 144,247,228,331,0 113,16,197,100,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000711.jpg 336,227,350,241,0 241,263,263,285,5 151,119,179,147,8 139,300,181,342,6 230,59,286,115,6 23,196,135,308,3 283,90,367,174,2 187,119,271,203,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000712.jpg 399,137,413,151,9 274,290,288,304,7 123,78,137,92,3 6,182,62,238,7 127,147,169,189,9 240,135,296,191,6 179,270,207,298,4 326,328,354,356,0 77,300,189,412,4 62,24,174,136,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000713.jpg 329,242,343,256,7 93,338,115,360,8 172,96,186,110,2 42,280,98,336,2 243,42,271,70,1 113,26,141,54,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000714.jpg 256,346,312,402,8 341,199,383,241,1 29,172,71,214,7 206,256,262,312,4 32,350,60,378,2 44,215,100,271,8 236,56,348,168,3 91,275,203,387,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000715.jpg 396,356,410,370,3 118,235,202,319,4 210,146,294,230,7 302,6,414,118,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000716.jpg 29,58,85,114,8 283,63,325,105,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000717.jpg 363,166,377,180,8 128,69,142,83,3 236,51,278,93,2 328,93,370,135,3 95,201,151,257,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000718.jpg 312,70,368,126,0 33,70,75,112,0 83,6,139,62,3 169,358,211,400,7 205,70,289,154,4 185,201,297,313,7 70,183,154,267,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000719.jpg 37,306,59,328,2 133,107,161,135,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000720.jpg 72,380,94,402,0 207,7,221,21,3 191,229,213,251,8 243,369,271,397,3 320,110,376,166,1 320,217,348,245,0 115,337,171,393,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000721.jpg 10,285,122,397,5 310,257,394,341,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000722.jpg 161,381,175,395,0 51,298,73,320,9 221,108,263,150,6 63,13,105,55,1 258,58,300,100,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000723.jpg 145,310,159,324,3 93,385,107,399,0 53,382,81,410,4 171,31,199,59,1 296,380,324,408,3 236,234,278,276,3 76,267,132,323,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000724.jpg 17,210,39,232,1 156,278,212,334,5 262,215,304,257,8 312,170,340,198,1 60,284,116,340,3 17,117,101,201,4 136,15,220,99,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000725.jpg 23,97,45,119,6 54,158,76,180,0 94,29,178,113,6 282,175,394,287,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000726.jpg 150,39,164,53,3 366,366,380,380,5 349,326,391,368,0 292,19,348,75,1 39,28,67,56,4 118,368,146,396,7 140,45,224,129,6 283,145,367,229,2 7,226,119,338,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000727.jpg 4,179,26,201,2 282,243,296,257,7 233,308,255,330,6 289,28,331,70,5 92,105,176,189,2 151,187,235,271,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000728.jpg 3,313,25,335,4 234,202,248,216,4 38,311,60,333,7 41,230,83,272,2 12,184,40,212,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000729.jpg 92,18,114,40,8 343,331,365,353,6 263,234,285,256,9 252,147,308,203,2 301,64,343,106,4 63,55,105,97,1 1,202,29,230,0 352,61,408,117,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000730.jpg 258,145,272,159,2 270,342,284,356,8 162,89,184,111,3 130,277,158,305,1 245,195,287,237,2 16,342,72,398,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000731.jpg 343,155,371,183,0 283,205,395,317,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000732.jpg 241,62,297,118,0 341,188,369,216,8 272,5,314,47,3 163,308,219,364,4 57,87,99,129,8 51,137,93,179,2 63,210,147,294,2 248,298,332,382,2 116,94,200,178,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000733.jpg 296,166,324,194,8 306,257,348,299,9 103,93,159,149,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000734.jpg 50,272,64,286,9 87,88,109,110,0 178,307,200,329,2 148,334,176,362,6 12,140,124,252,9 163,117,247,201,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000735.jpg 59,76,87,104,9 189,128,245,184,4 32,232,144,344,5 287,103,399,215,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000736.jpg 306,357,320,371,2 90,397,104,411,9 252,307,294,349,3 33,273,75,315,6 278,369,320,411,0 128,375,156,403,2 77,149,189,261,7 91,35,203,147,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000737.jpg 123,70,151,98,8 262,48,304,90,9 5,148,61,204,6 367,114,409,156,4 251,326,293,368,5 89,210,201,322,3 14,330,98,414,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000738.jpg 64,401,78,415,8 236,85,278,127,9 336,340,378,382,8 139,132,181,174,2 108,344,164,400,4 196,172,238,214,7 373,30,415,72,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000739.jpg 169,173,211,215,2 55,275,167,387,6 241,63,353,175,1 280,235,364,319,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000740.jpg 169,361,211,403,4 17,301,59,343,3 0,64,42,106,7 170,259,254,343,6 73,92,185,204,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000741.jpg 89,272,111,294,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000742.jpg 226,100,282,156,6 354,355,382,383,1 266,68,294,96,5 371,111,413,153,8 174,186,230,242,3 151,128,179,156,9 275,230,387,342,0 49,238,133,322,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000743.jpg 298,51,354,107,9 39,65,151,177,6 177,29,289,141,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000744.jpg 279,299,293,313,8 132,75,160,103,0 233,87,289,143,9 69,168,181,280,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000745.jpg 203,239,225,261,1 112,35,134,57,2 109,132,165,188,9 176,43,232,99,5 93,233,149,289,7 64,332,106,374,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000746.jpg 287,112,309,134,9 330,39,344,53,0 171,63,283,175,2 33,78,117,162,7 251,286,335,370,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000747.jpg 194,45,306,157,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000748.jpg 78,284,100,306,7 39,322,53,336,2 322,329,336,343,2 116,207,172,263,5 285,381,313,409,8 119,374,147,402,8 344,328,400,384,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000749.jpg 327,111,349,133,4 346,12,368,34,3 34,244,48,258,0 82,330,110,358,4 317,312,345,340,8 157,273,185,301,4 257,28,313,84,4 68,162,96,190,5 212,197,324,309,6 121,194,205,278,5 146,58,258,170,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000750.jpg 2,47,24,69,8 73,73,185,185,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000751.jpg 193,197,221,225,5 151,4,193,46,3 259,57,287,85,7 353,154,381,182,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000752.jpg 240,55,254,69,0 41,386,55,400,6 143,374,171,402,1 190,177,218,205,0 70,350,98,378,7 39,131,81,173,4 251,158,363,270,6 8,262,92,346,7 249,42,333,126,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000753.jpg 378,256,392,270,5 213,222,227,236,8 252,196,294,238,1 178,130,220,172,9 24,346,52,374,4 299,243,411,355,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000754.jpg 91,191,133,233,9 78,273,134,329,4 316,116,344,144,9 213,150,325,262,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000755.jpg 319,322,375,378,7 183,222,211,250,0 309,36,351,78,9 53,197,109,253,6 331,104,373,146,9 371,258,413,300,3 48,299,160,411,2 225,9,309,93,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000756.jpg 380,337,402,359,4 313,89,327,103,2 202,389,216,403,6 338,107,366,135,5 0,162,28,190,2 68,37,96,65,9 268,28,296,56,6 144,238,228,322,1 124,67,208,151,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000757.jpg 119,351,141,373,1 11,282,33,304,0 110,189,166,245,1 202,239,244,281,3 252,208,364,320,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000758.jpg 284,370,306,392,0 377,288,391,302,1 157,61,269,173,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000759.jpg 143,179,185,221,0 229,239,285,295,1 234,15,346,127,3 276,290,388,402,3 258,135,342,219,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000760.jpg 343,198,365,220,1 88,121,116,149,1 117,162,159,204,3 178,323,220,365,4 197,14,239,56,9 220,297,262,339,4 166,136,278,248,0 270,276,354,360,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000761.jpg 173,306,229,362,0 356,298,412,354,3 110,353,166,409,2 208,121,320,233,7 67,201,179,313,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000762.jpg 92,172,120,200,0 256,121,284,149,6 21,335,49,363,7 287,114,315,142,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000763.jpg 26,0,82,56,7 31,261,59,289,3 146,141,258,253,1 126,26,210,110,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000764.jpg 105,3,147,45,3 147,241,259,353,2 218,82,302,166,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000765.jpg 0,365,14,379,5 98,293,210,405,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000766.jpg 253,53,267,67,5 372,85,386,99,3 352,261,374,283,8 166,159,222,215,9 134,101,176,143,3 21,1,49,29,9 188,311,244,367,0 227,221,269,263,5 53,76,81,104,2 193,31,305,143,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000767.jpg 397,201,411,215,6 368,147,382,161,6 309,175,331,197,3 274,359,302,387,9 222,275,250,303,7 142,220,198,276,4 309,20,365,76,8 200,135,242,177,1 52,273,136,357,9 21,175,105,259,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000768.jpg 81,360,95,374,6 390,56,412,78,2 185,31,227,73,3 235,23,319,107,3 78,197,190,309,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000769.jpg 153,272,209,328,6 269,134,381,246,8 255,253,339,337,9 108,65,192,149,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000770.jpg 153,150,175,172,3 26,39,40,53,9 368,133,382,147,4 213,302,297,386,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000771.jpg 107,336,129,358,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000772.jpg 292,113,348,169,3 18,312,74,368,0 1,273,29,301,0 149,81,177,109,9 126,164,210,248,6 234,267,346,379,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000773.jpg 72,235,114,277,8 246,277,274,305,4 249,26,291,68,4 317,110,373,166,4 227,79,283,135,7 30,102,114,186,9 236,309,320,393,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000774.jpg 120,154,134,168,3 280,123,336,179,0 252,301,280,329,9 169,108,281,220,6 68,288,180,400,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000775.jpg 260,23,274,37,7 299,253,313,267,6 278,137,320,179,1 109,241,221,353,0 126,153,210,237,5 57,9,169,121,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000776.jpg 152,129,174,151,5 306,272,320,286,6 247,204,303,260,0 124,120,152,148,3 276,86,304,114,6 102,27,130,55,8 123,245,235,357,1 275,264,387,376,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000777.jpg 392,342,406,356,3 123,49,165,91,6 243,205,327,289,7 201,53,313,165,6 63,106,147,190,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000778.jpg 292,219,314,241,1 109,227,137,255,9 170,291,254,375,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000779.jpg 263,159,285,181,2 92,45,120,73,9 315,1,343,29,1 22,305,50,333,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000780.jpg 345,249,367,271,1 265,43,293,71,8 319,46,375,102,9 203,231,245,273,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000781.jpg 225,181,239,195,1 274,58,330,114,5 83,117,139,173,2 364,99,406,141,2 24,59,52,87,3 253,118,365,230,4 296,244,408,356,6 23,169,135,281,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000782.jpg 201,141,215,155,3 307,308,329,330,9 210,8,238,36,9 317,209,401,293,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000783.jpg 225,110,239,124,8 73,44,95,66,5 6,40,28,62,4 135,118,191,174,4 0,371,42,413,1 321,35,349,63,2 217,221,259,263,7 130,44,186,100,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000784.jpg 162,9,274,121,7 3,106,115,218,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000785.jpg 138,231,180,273,2 221,382,249,410,2 235,240,291,296,1 106,209,134,237,3 284,150,340,206,7 4,309,60,365,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000786.jpg 174,257,202,285,3 220,45,262,87,0 63,325,105,367,0 216,185,244,213,5 149,121,205,177,5 131,65,159,93,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000787.jpg 289,157,311,179,5 298,338,312,352,7 341,80,363,102,1 139,34,167,62,1 91,148,119,176,4 271,209,383,321,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000788.jpg 271,93,285,107,5 347,330,369,352,6 168,96,182,110,2 136,380,164,408,2 167,310,209,352,6 220,241,276,297,9 327,181,369,223,5 288,290,316,318,8 19,209,131,321,0 42,60,154,172,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000789.jpg 93,5,115,27,3 367,327,395,355,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000790.jpg 149,32,163,46,3 4,261,18,275,7 189,297,245,353,4 332,119,388,175,3 259,69,315,125,3 325,233,353,261,3 277,327,305,355,2 58,283,142,367,8 135,43,247,155,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000791.jpg 69,195,125,251,6 227,328,255,356,2 229,61,285,117,3 169,43,225,99,0 250,164,278,192,5 262,272,374,384,3 59,21,143,105,7 38,278,122,362,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000792.jpg 356,40,378,62,5 401,184,415,198,5 104,174,126,196,1 294,22,336,64,8 60,85,88,113,3 263,176,291,204,1 66,296,178,408,1 292,287,376,371,3 153,98,265,210,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000793.jpg 282,154,324,196,5 56,18,98,60,2 157,135,241,219,0 1,132,113,244,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000794.jpg 192,68,206,82,4 323,328,337,342,4 44,85,58,99,0 184,212,240,268,5 67,304,109,346,0 344,290,372,318,9 19,201,103,285,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000795.jpg 267,342,295,370,1 156,387,184,415,6 29,94,85,150,2 95,20,151,76,6 86,154,114,182,3 17,188,73,244,1 155,1,239,85,3 224,246,308,330,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000796.jpg 351,254,379,282,3 158,97,200,139,7 15,155,71,211,1 148,156,204,212,8 236,264,292,320,0 242,93,284,135,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000797.jpg 153,328,175,350,9 318,5,360,47,7 383,90,411,118,9 174,101,286,213,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000798.jpg 369,38,383,52,4 135,220,219,304,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000799.jpg 197,179,211,193,8 77,268,133,324,2 278,133,320,175,4 146,270,188,312,8 20,45,48,73,9 33,189,89,245,2 58,134,86,162,2 321,5,405,89,7 282,204,394,316,2 113,77,225,189,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000800.jpg 69,330,91,352,0 191,258,303,370,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000801.jpg 200,206,214,220,1 254,122,310,178,7 261,331,303,373,1 168,59,210,101,5 282,15,324,57,9 303,297,331,325,2 42,51,70,79,5 87,250,199,362,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000802.jpg 3,374,17,388,1 233,299,275,341,1 363,316,405,358,1 37,312,93,368,9 198,257,226,285,1 256,215,312,271,5 25,193,67,235,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000803.jpg 107,105,129,127,9 54,2,68,16,5 87,152,109,174,4 269,124,325,180,1 187,173,229,215,8 12,261,124,373,0 295,10,379,94,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000804.jpg 100,180,122,202,2 153,110,167,124,2 45,181,59,195,4 215,312,257,354,3 210,372,252,414,9 241,187,297,243,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000805.jpg 172,274,186,288,1 278,195,300,217,7 58,267,72,281,0 308,350,364,406,9 22,333,64,375,2 54,13,82,41,6 261,10,317,66,4 199,8,227,36,2 26,41,82,97,2 243,211,355,323,4 119,1,203,85,0 244,92,328,176,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000806.jpg 377,216,399,238,5 399,364,413,378,5 228,301,270,343,3 30,264,142,376,7 221,23,305,107,2 302,31,414,143,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000807.jpg 289,333,317,361,1 335,70,391,126,3 331,228,373,270,7 182,310,238,366,8 201,67,257,123,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000808.jpg 77,261,119,303,6 49,236,77,264,9 317,178,373,234,8 2,22,44,64,3 156,369,184,397,5 168,260,280,372,1 255,9,367,121,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000809.jpg 365,151,379,165,8 97,314,119,336,7 52,355,74,377,2 367,16,409,58,3 214,279,256,321,5 178,76,206,104,9 60,2,102,44,8 31,112,143,224,9 286,59,370,143,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000810.jpg 70,247,84,261,7 228,67,242,81,7 62,154,104,196,1 133,343,175,385,0 331,243,387,299,2 255,251,297,293,8 92,259,148,315,7 104,121,216,233,9 269,147,353,231,6 200,1,312,113,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000811.jpg 272,377,286,391,1 213,236,235,258,1 62,255,104,297,5 200,283,256,339,7 19,272,61,314,4 358,238,414,294,2 215,86,327,198,1 3,49,115,161,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000812.jpg 166,264,180,278,5 31,111,73,153,9 183,167,239,223,7 38,350,80,392,9 49,56,105,112,4 242,292,326,376,3 256,177,340,261,4 101,6,213,118,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000813.jpg 350,220,364,234,1 297,298,353,354,7 269,289,297,317,6 82,258,110,286,3 68,154,180,266,7 211,56,295,140,0 111,66,195,150,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000814.jpg 71,197,99,225,4 320,185,348,213,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000815.jpg 267,160,289,182,7 247,296,261,310,8 10,62,52,104,8 68,286,180,398,9 287,4,399,116,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000816.jpg 360,296,374,310,0 344,232,358,246,7 153,213,237,297,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000817.jpg 233,128,247,142,5 305,230,327,252,1 143,242,157,256,9 168,181,210,223,0 297,305,325,333,8 129,281,241,393,1 22,67,134,179,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000818.jpg 99,381,113,395,8 27,79,83,135,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000819.jpg 373,142,401,170,5 52,333,108,389,2 189,224,245,280,5 250,120,292,162,4 85,98,197,210,3 301,293,413,405,4 104,253,188,337,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000820.jpg 263,107,285,129,0 295,33,337,75,4 89,40,131,82,9 366,218,408,260,1 145,233,201,289,8 199,107,255,163,3 75,298,159,382,0 252,192,336,276,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000821.jpg 271,26,285,40,6 239,138,253,152,4 194,69,236,111,9 152,125,180,153,0 151,212,179,240,0 327,304,355,332,2 236,278,292,334,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000822.jpg 187,96,243,152,1 354,297,410,353,1 103,80,159,136,3 178,307,206,335,6 93,149,205,261,9 293,138,405,250,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000823.jpg 348,175,362,189,0 22,45,36,59,8 305,213,319,227,2 155,49,211,105,0 313,117,341,145,5 34,172,62,200,0 262,297,304,339,6 61,165,145,249,2 312,330,396,414,3 218,15,330,127,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000824.jpg 18,0,40,22,8 113,293,225,405,7 69,162,153,246,0 314,211,398,295,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000825.jpg 59,333,115,389,1 310,324,352,366,8 140,105,168,133,4 328,104,384,160,5 258,61,314,117,9 133,271,189,327,5 24,25,136,137,9 39,136,123,220,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000826.jpg 86,50,100,64,0 193,275,305,387,0 315,26,399,110,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000827.jpg 32,183,54,205,1 279,271,321,313,9 300,64,342,106,3 22,56,78,112,2 225,60,253,88,8 95,113,151,169,0 255,101,311,157,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000828.jpg 320,119,342,141,9 106,35,120,49,9 265,89,307,131,4 131,116,173,158,4 238,290,280,332,3 255,11,283,39,6 174,115,216,157,6 344,99,372,127,0 268,190,352,274,5 178,33,262,117,3 86,267,170,351,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000829.jpg 162,151,176,165,1 172,379,194,401,7 42,189,64,211,1 2,14,58,70,0 180,189,236,245,7 234,188,276,230,5 103,123,159,179,4 195,23,237,65,4 253,221,365,333,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000830.jpg 214,347,228,361,7 76,127,98,149,1 243,228,271,256,1 345,118,401,174,8 341,46,397,102,0 165,123,221,179,6 119,8,161,50,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000831.jpg 276,104,290,118,3 364,8,386,30,0 86,331,100,345,0 30,54,114,138,3 188,127,300,239,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000832.jpg 193,5,249,61,2 323,158,365,200,6 278,68,306,96,4 260,121,316,177,3 171,274,283,386,4 67,7,151,91,7 236,177,320,261,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000833.jpg 11,127,25,141,7 132,100,154,122,0 316,134,372,190,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000834.jpg 300,170,342,212,9 381,113,409,141,9 93,336,149,392,5 237,326,321,410,4 164,188,276,300,7 284,0,368,84,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000835.jpg 223,241,245,263,5 12,244,26,258,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000836.jpg 81,105,95,119,7 341,307,363,329,9 70,173,84,187,8 105,202,133,230,6 321,24,349,52,4 294,354,350,410,8 74,66,102,94,0 189,202,217,230,0 81,22,123,64,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000837.jpg 113,128,135,150,2 373,154,395,176,6 110,51,124,65,1 329,209,385,265,1 119,276,161,318,3 161,212,189,240,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000838.jpg 25,139,67,181,9 292,190,404,302,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000839.jpg 381,291,395,305,4 65,165,87,187,8 237,40,259,62,8 241,62,283,104,7 191,86,219,114,3 373,307,415,349,8 13,134,41,162,3 320,3,404,87,0 221,105,333,217,7 8,252,120,364,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000840.jpg 306,214,320,228,6 6,371,20,385,9 87,293,171,377,0 300,241,412,353,5 168,74,252,158,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000841.jpg 363,152,377,166,7 176,334,232,390,4 19,138,61,180,8 246,18,274,46,2 40,261,96,317,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000842.jpg 198,279,220,301,9 254,286,366,398,9 70,283,182,395,0 224,61,308,145,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000843.jpg 96,299,152,355,8 240,261,268,289,8 257,68,313,124,2 238,41,266,69,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000844.jpg 197,20,219,42,1 149,257,163,271,8 351,163,373,185,6 275,41,303,69,5 316,248,372,304,7 121,363,149,391,0 250,262,278,290,5 72,174,156,258,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000845.jpg 203,319,225,341,7 306,5,362,61,3 222,335,264,377,1 251,250,307,306,2 97,176,181,260,0 270,120,354,204,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000846.jpg 146,48,168,70,8 348,17,404,73,0 278,90,320,132,8 269,339,297,367,1 4,199,32,227,1 228,225,270,267,1 16,347,58,389,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000847.jpg 362,35,376,49,1 211,333,233,355,8 147,161,189,203,0 25,75,53,103,3 222,301,334,413,7 75,326,159,410,7 58,177,142,261,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000848.jpg 317,325,331,339,2 188,219,202,233,9 147,304,189,346,4 218,140,274,196,5 205,357,261,413,7 112,8,168,64,2 293,82,377,166,1 84,69,196,181,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000849.jpg 226,163,248,185,0 294,140,336,182,5 329,251,385,307,8 93,74,121,102,3 112,143,140,171,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000850.jpg 242,375,256,389,3 277,386,291,400,6 366,81,388,103,1 180,269,208,297,4 2,107,58,163,2 106,91,148,133,2 262,79,304,121,4 149,334,205,390,3 22,182,78,238,7 101,214,185,298,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000851.jpg 283,117,311,145,8 168,276,224,332,1 333,353,361,381,7 309,192,365,248,5 197,142,309,254,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000852.jpg 395,355,409,369,5 202,33,216,47,1 182,67,224,109,2 315,126,343,154,0 310,169,366,225,6 235,270,263,298,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000853.jpg 238,75,294,131,5 327,324,369,366,1 13,330,69,386,2 265,129,377,241,1 88,158,172,242,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000854.jpg 99,66,121,88,5 9,328,37,356,0 192,169,304,281,3 115,316,199,400,0 228,283,312,367,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000855.jpg 5,103,27,125,7 288,145,344,201,3 350,231,378,259,0 88,300,200,412,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000856.jpg 262,118,276,132,7 236,228,278,270,5 153,250,209,306,7 30,296,114,380,4 53,50,165,162,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000857.jpg 248,237,270,259,9 13,299,35,321,6 154,229,196,271,8 37,232,93,288,2 149,319,177,347,8 176,273,260,357,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000858.jpg 369,194,397,222,1 308,325,364,381,8 359,67,401,109,4 223,346,265,388,9 271,101,383,213,6 83,42,195,154,3 316,223,400,307,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000859.jpg 262,77,276,91,1 56,168,78,190,8 332,108,360,136,9 176,80,288,192,3 104,258,216,370,8 279,251,363,335,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000860.jpg 101,306,123,328,6 346,374,360,388,0 24,298,66,340,1 16,46,44,74,2 156,200,212,256,2 296,58,338,100,5 235,202,347,314,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000861.jpg 345,195,359,209,1 3,242,17,256,3 285,359,327,401,7 154,122,182,150,5 94,241,150,297,6 255,44,297,86,3 217,287,245,315,4 30,97,72,139,8 278,111,362,195,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000862.jpg 285,124,307,146,5 126,8,140,22,8 40,62,54,76,3 279,263,335,319,0 332,230,374,272,4 320,45,362,87,5 337,108,379,150,1 240,321,268,349,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000863.jpg 122,13,136,27,1 13,64,69,120,9 226,172,254,200,1 260,256,288,284,9 182,157,224,199,6 341,16,369,44,1 84,295,196,407,0 219,308,303,392,4 200,80,284,164,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000864.jpg 284,81,298,95,1 222,304,244,326,1 116,213,200,297,7 263,250,375,362,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000865.jpg 268,105,290,127,9 20,64,34,78,7 141,83,183,125,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000866.jpg 10,67,94,151,6 278,76,362,160,8 243,236,355,348,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000867.jpg 193,198,207,212,0 55,138,77,160,9 354,219,376,241,9 334,36,390,92,1 47,310,75,338,4 242,114,354,226,1 259,259,371,371,4 133,28,245,140,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000868.jpg 120,192,142,214,5 312,190,354,232,8 371,77,399,105,9 224,148,280,204,3 11,5,67,61,3 323,238,351,266,9 175,157,203,185,6 191,291,275,375,9 0,292,112,404,4 111,331,195,415,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000869.jpg 285,248,299,262,3 339,202,353,216,1 15,87,57,129,7 137,195,165,223,6 326,261,410,345,8 18,295,130,407,8 176,60,260,144,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000870.jpg 336,164,350,178,9 194,92,236,134,2 14,299,70,355,7 228,174,256,202,3 166,275,222,331,6 363,9,405,51,3 296,253,324,281,5 38,193,150,305,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000871.jpg 100,109,114,123,7 203,109,225,131,6 337,44,393,100,2 113,238,225,350,3 256,287,368,399,3 124,19,208,103,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000872.jpg 0,223,28,251,2 269,7,381,119,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000873.jpg 118,334,132,348,9 128,121,184,177,3 116,192,172,248,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000874.jpg 346,244,368,266,2 218,308,232,322,1 134,271,148,285,3 30,76,72,118,3 55,281,111,337,7 117,138,173,194,0 204,116,288,200,2 301,68,413,180,9 204,209,316,321,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000875.jpg 260,320,274,334,4 115,63,137,85,6 69,356,83,370,5 48,267,90,309,7 164,116,276,228,0 281,297,365,381,5 113,253,225,365,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000876.jpg 238,373,252,387,7 26,363,48,385,7 125,58,181,114,5 113,301,155,343,1 227,190,269,232,7 34,86,62,114,9 295,268,379,352,4 258,87,342,171,1 17,172,129,284,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000877.jpg 40,237,54,251,2 247,317,269,339,5 366,21,394,49,3 117,91,229,203,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000878.jpg 50,31,72,53,3 105,233,147,275,9 298,96,382,180,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000879.jpg 221,277,243,299,8 364,258,378,272,1 165,83,179,97,9 224,41,266,83,0 91,283,133,325,5 355,168,397,210,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000880.jpg 293,89,321,117,2 337,93,379,135,7 353,154,395,196,8 266,336,322,392,8 89,44,131,86,8 209,123,237,151,5 186,29,270,113,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000881.jpg 117,100,131,114,4 183,379,205,401,7 356,118,384,146,9 379,69,407,97,9 248,170,276,198,6 232,43,288,99,4 243,347,285,389,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000882.jpg 294,114,316,136,9 231,177,253,199,6 354,85,376,107,0 76,155,104,183,1 95,190,137,232,9 385,310,413,338,5 28,51,140,163,7 161,281,245,365,0 239,16,323,100,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000883.jpg 352,209,374,231,3 378,123,400,145,9 226,56,254,84,9 44,154,100,210,9 119,343,147,371,6 65,36,149,120,5 93,193,205,305,0 302,311,386,395,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000884.jpg 184,12,206,34,2 174,55,216,97,3 261,317,289,345,8 104,130,146,172,1 209,77,293,161,6 132,231,244,343,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000885.jpg 161,329,189,357,8 136,17,178,59,2 186,131,242,187,9 86,265,142,321,5 112,358,140,386,6 327,302,369,344,1 29,42,141,154,6 267,167,379,279,2 191,232,275,316,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000886.jpg 239,173,253,187,5 265,9,279,23,8 321,267,363,309,1 223,79,279,135,0 310,154,338,182,4 131,349,173,391,5 368,102,410,144,1 66,162,178,274,4 88,36,172,120,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000887.jpg 62,226,84,248,6 32,366,46,380,8 242,325,264,347,6 307,21,363,77,1 133,58,189,114,5 54,46,110,102,8 20,164,48,192,7 52,298,108,354,2 157,314,199,356,9 290,224,374,308,4 73,146,185,258,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000888.jpg 88,93,130,135,0 162,61,218,117,4 21,212,77,268,0 222,258,250,286,1 333,208,361,236,3 234,171,318,255,2 73,272,185,384,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000889.jpg 285,344,299,358,6 294,380,308,394,4 140,239,182,281,0 378,173,406,201,4 322,197,364,239,2 306,243,334,271,0 288,40,372,124,1 72,128,184,240,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000890.jpg 288,67,372,151,8 157,33,269,145,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000891.jpg 64,12,78,26,1 270,164,284,178,4 254,36,310,92,8 186,362,228,404,9 303,150,359,206,1 53,299,137,383,8 58,134,170,246,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000892.jpg 223,101,237,115,0 382,81,410,109,9 134,34,176,76,8 5,136,33,164,4 230,104,314,188,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000893.jpg 158,68,172,82,7 160,219,174,233,8 256,210,298,252,0 211,12,239,40,8 68,331,152,415,6 310,226,394,310,6 72,156,156,240,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000894.jpg 152,32,166,46,1 390,104,412,126,2 57,293,71,307,7 192,50,248,106,3 86,119,128,161,2 38,119,66,147,6 223,108,251,136,8 133,321,175,363,1 344,220,372,248,1 19,264,131,376,8 228,188,312,272,3 139,186,223,270,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000895.jpg 121,342,143,364,7 246,242,288,284,0 345,15,401,71,1 324,121,408,205,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000896.jpg 104,44,118,58,4 269,3,297,31,8 166,34,208,76,7 299,244,383,328,6 106,160,190,244,6 195,302,279,386,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000897.jpg 258,298,280,320,1 51,105,65,119,4 105,199,161,255,1 188,217,230,259,0 136,312,192,368,5 66,47,178,159,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000898.jpg 232,159,274,201,1 262,354,304,396,2 374,221,402,249,1 239,295,281,337,3 120,141,232,253,0 106,24,218,136,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000899.jpg 183,95,211,123,5 223,116,251,144,1 299,39,341,81,5 96,32,124,60,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000900.jpg 117,377,139,399,3 299,179,313,193,8 179,141,221,183,9 118,332,160,374,0 163,368,205,410,1 101,98,129,126,3 23,14,79,70,3 230,13,314,97,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000901.jpg 64,385,78,399,9 320,272,362,314,7 178,298,234,354,2 117,0,159,42,7 280,201,336,257,6 86,45,170,129,2 68,228,180,340,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000902.jpg 208,337,230,359,1 166,301,208,343,0 126,127,182,183,2 274,353,302,381,1 16,225,72,281,2 279,156,321,198,7 251,226,335,310,5 143,221,227,305,1 75,308,159,392,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000903.jpg 274,369,288,383,5 356,401,370,415,5 46,352,68,374,3 55,171,83,199,9 334,279,390,335,1 239,16,267,44,0 36,94,92,150,1 92,213,148,269,0 115,358,157,400,8 259,153,343,237,3 222,293,334,405,7 142,142,254,254,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000904.jpg 165,179,207,221,0 14,207,42,235,7 256,66,284,94,2 266,128,350,212,8 109,257,221,369,3 271,245,355,329,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000905.jpg 283,333,297,347,8 32,292,88,348,9 377,283,405,311,2 373,123,401,151,0 189,33,217,61,9 330,374,358,402,2 221,364,249,392,5 35,102,147,214,1 237,238,349,350,2 143,272,227,356,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000906.jpg 169,57,183,71,0 234,172,276,214,9 0,191,28,219,9 346,194,388,236,2 34,188,146,300,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000907.jpg 229,313,251,335,1 244,80,258,94,0 380,56,402,78,4 75,0,103,28,2 60,77,88,105,0 131,281,187,337,1 50,316,78,344,8 192,104,220,132,0 291,247,403,359,7 55,100,139,184,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000908.jpg 158,94,180,116,6 364,368,378,382,7 216,71,272,127,4 251,201,293,243,7 5,89,61,145,6 227,11,255,39,9 281,329,365,413,8 77,218,189,330,0 152,133,236,217,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000909.jpg 34,97,56,119,4 396,152,410,166,1 248,146,270,168,0 99,140,183,224,4 108,38,192,122,6 32,241,116,325,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000910.jpg 345,241,367,263,5 380,188,402,210,5 122,236,144,258,8 212,363,254,405,4 28,15,70,57,6 156,268,240,352,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000911.jpg 323,22,345,44,9 372,114,386,128,3 85,235,107,257,9 110,51,152,93,5 40,126,152,238,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000912.jpg 164,300,220,356,6 305,238,333,266,1 295,178,337,220,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000913.jpg 89,170,103,184,6 309,158,323,172,3 21,187,35,201,7 354,250,382,278,7 335,30,391,86,8 56,240,84,268,3 98,310,126,338,2 348,311,376,339,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000914.jpg 343,350,365,372,2 19,308,47,336,4 246,275,274,303,8 378,279,406,307,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000915.jpg 224,95,246,117,0 291,220,313,242,9 300,28,412,140,5 66,166,178,278,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000916.jpg 303,180,345,222,3 290,304,318,332,7 318,57,374,113,8 217,5,259,47,2 106,322,134,350,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000917.jpg 245,53,259,67,2 283,220,305,242,9 42,340,98,396,5 45,116,101,172,6 331,228,359,256,5 90,264,174,348,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000918.jpg 356,215,378,237,7 220,160,242,182,4 250,194,264,208,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000919.jpg 48,50,132,134,4 81,160,165,244,9 228,263,340,375,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000920.jpg 205,323,219,337,1 178,108,200,130,7 132,27,174,69,4 34,143,90,199,1 243,294,271,322,0 16,56,100,140,9 263,145,375,257,2 265,23,349,107,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000921.jpg 324,164,380,220,4 131,237,173,279,8 298,23,340,65,7 237,7,279,49,1 127,8,183,64,9 261,264,345,348,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000922.jpg 54,138,68,152,5 56,259,78,281,5 97,267,119,289,4 216,143,244,171,0 249,270,305,326,5 288,195,330,237,5 223,297,251,325,9 56,24,168,136,1 302,296,414,408,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000923.jpg 151,32,165,46,1 205,295,247,337,7 103,74,131,102,0 150,189,192,231,9 43,242,85,284,8 272,25,328,81,9 338,225,380,267,2 11,101,95,185,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000924.jpg 202,120,224,142,3 110,21,166,77,2 48,298,90,340,3 265,36,349,120,6 181,247,293,359,1 46,177,158,289,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000925.jpg 322,195,336,209,9 379,103,401,125,4 85,167,127,209,2 51,64,93,106,5 349,166,377,194,6 153,310,237,394,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000926.jpg 333,101,355,123,1 13,371,41,399,7 65,35,121,91,7 11,215,39,243,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000927.jpg 73,140,87,154,9 146,44,168,66,1 16,261,30,275,7 117,245,173,301,4 132,330,188,386,7 323,342,351,370,1 58,45,114,101,7 305,194,389,278,1 29,230,113,314,3 112,55,224,167,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000928.jpg 336,69,358,91,9 367,155,389,177,6 106,143,148,185,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000929.jpg 269,56,283,70,0 358,245,380,267,0 106,296,148,338,3 77,166,133,222,2 272,252,300,280,0 316,227,344,255,5 43,297,99,353,1 53,36,137,120,7 161,180,273,292,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000930.jpg 182,308,204,330,8 179,73,201,95,5 96,143,124,171,2 339,281,367,309,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000931.jpg 350,218,372,240,2 51,197,79,225,6 160,289,202,331,8 133,211,189,267,7 332,155,374,197,2 268,243,310,285,6 297,21,409,133,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000932.jpg 149,21,171,43,7 24,66,38,80,1 300,27,412,139,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000933.jpg 28,268,50,290,2 264,180,306,222,3 370,374,398,402,5 92,257,134,299,9 158,92,200,134,5 369,269,411,311,9 36,123,148,235,1 180,313,264,397,3 49,40,133,124,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000934.jpg 236,300,292,356,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000935.jpg 66,255,80,269,1 135,229,191,285,3 367,1,409,43,5 337,259,365,287,2 71,319,127,375,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000936.jpg 65,44,149,128,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000937.jpg 81,337,95,351,9 246,349,274,377,8 325,21,367,63,3 196,262,280,346,8 152,18,264,130,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000938.jpg 246,72,268,94,5 218,13,240,35,6 115,275,137,297,7 159,132,187,160,4 312,178,354,220,9 157,88,199,130,9 280,116,336,172,1 32,353,60,381,0 116,0,158,42,3 165,298,277,410,2 322,245,406,329,1 4,148,88,232,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000939.jpg 61,260,89,288,7 101,318,129,346,1 158,224,214,280,8 35,76,63,104,2 226,153,254,181,7 31,181,87,237,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000940.jpg 387,73,415,101,4 358,10,400,52,3 267,71,309,113,0 54,51,110,107,3 144,114,256,226,4 58,198,142,282,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000941.jpg 329,327,357,355,3 277,340,333,396,0 243,141,299,197,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000942.jpg 232,199,246,213,9 215,349,237,371,8 340,70,362,92,9 76,204,104,232,4 127,227,155,255,5 355,135,397,177,9 242,208,354,320,1 71,90,183,202,3 168,5,252,89,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000943.jpg 42,279,64,301,1 266,327,280,341,4 217,107,239,129,2 196,359,252,415,2 267,291,379,403,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000944.jpg 349,197,371,219,0 3,117,17,131,3 257,306,299,348,7 182,78,224,120,4 43,195,85,237,0 124,230,166,272,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000945.jpg 12,284,26,298,0 281,366,323,408,6 23,21,65,63,0 172,57,214,99,1 2,287,114,399,5 25,136,109,220,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000946.jpg 38,79,52,93,5 247,140,303,196,5 345,6,401,62,7 207,72,249,114,8 163,241,191,269,5 322,59,378,115,4 5,202,117,314,2 100,117,212,229,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000947.jpg 23,91,37,105,1 200,281,214,295,4 344,62,372,90,3 38,363,80,405,3 174,28,230,84,4 243,104,271,132,7 101,274,185,358,1 228,315,312,399,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000948.jpg 306,352,320,366,3 374,4,396,26,8 121,236,135,250,0 279,60,307,88,0 3,320,59,376,0 198,133,226,161,2 175,170,217,212,2 261,241,345,325,0 31,186,143,298,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000949.jpg 65,318,79,332,4 0,382,14,396,0 307,206,329,228,9 81,323,109,351,7 295,292,407,404,5 28,203,140,315,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000950.jpg 67,338,89,360,2 60,358,116,414,7 145,149,201,205,6 84,132,112,160,1 297,57,325,85,8 110,1,166,57,1 140,238,224,322,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000951.jpg 102,6,144,48,1 68,309,96,337,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000952.jpg 167,91,223,147,9 107,204,149,246,5 361,279,403,321,5 184,233,212,261,3 146,16,202,72,4 176,272,260,356,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000953.jpg 59,381,81,403,7 84,147,126,189,6 231,298,273,340,7 7,219,91,303,2 222,95,334,207,7 218,0,302,84,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000954.jpg 11,109,67,165,9 163,146,247,230,8 277,155,389,267,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000955.jpg 199,348,221,370,5 220,275,242,297,7 51,362,93,404,6 18,358,46,386,1 63,73,105,115,6 94,257,178,341,4 111,28,223,140,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000956.jpg 342,7,364,29,8 36,205,58,227,8 108,356,130,378,1 277,156,333,212,8 320,78,362,120,7 159,97,201,139,9 24,152,80,208,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000957.jpg 174,96,216,138,2 199,196,255,252,3 158,345,186,373,6 301,12,357,68,9 60,281,102,323,3 24,147,136,259,9 257,95,369,207,4 330,266,414,350,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000958.jpg 210,68,224,82,1 298,19,340,61,4 239,248,267,276,2 113,298,169,354,6 21,269,77,325,5 173,277,201,305,9 245,94,301,150,3 290,254,402,366,8 82,57,194,169,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000959.jpg 51,210,73,232,7 207,362,235,390,6 15,11,57,53,9 223,302,265,344,9 276,279,318,321,7 68,131,152,215,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000960.jpg 209,162,231,184,8 37,23,51,37,5 107,99,129,121,5 301,39,329,67,5 26,143,82,199,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000961.jpg 271,234,293,256,3 277,144,299,166,7 216,278,230,292,7 276,290,304,318,6 185,125,213,153,4 75,148,131,204,8 313,208,341,236,8 91,272,147,328,1 305,89,389,173,6 310,245,394,329,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000962.jpg 171,352,185,366,7 279,95,301,117,4 21,129,43,151,4 17,33,73,89,8 42,216,98,272,7 217,267,245,295,1 166,192,194,220,0 322,49,350,77,7 105,295,217,407,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000963.jpg 312,22,396,106,4 66,315,150,399,0 129,88,241,200,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000964.jpg 313,259,327,273,1 92,131,148,187,7 358,123,400,165,6 26,341,54,369,2 205,53,247,95,2 288,161,344,217,5 15,243,57,285,1 152,277,264,389,0 58,236,142,320,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000965.jpg 92,125,106,139,9 290,369,312,391,5 344,8,358,22,2 295,201,337,243,2 0,40,28,68,0 81,57,123,99,6 336,192,364,220,9 285,71,313,99,2 103,28,131,56,5 181,279,265,363,2 157,124,269,236,2 300,265,384,349,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000966.jpg 131,223,173,265,1 328,63,356,91,6 259,128,371,240,2 145,18,257,130,1 66,286,178,398,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000967.jpg 66,300,88,322,3 5,301,33,329,3 52,84,94,126,7 95,265,179,349,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000968.jpg 2,354,58,410,7 128,147,184,203,3 302,214,344,256,4 225,56,267,98,1 270,160,312,202,2 264,266,306,308,6 32,48,144,160,2 118,295,202,379,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000969.jpg 377,267,405,295,9 213,311,255,353,7 100,103,156,159,9 269,271,353,355,2 229,189,313,273,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000970.jpg 123,174,137,188,5 234,138,256,160,6 131,321,153,343,4 221,245,263,287,6 114,184,226,296,1 300,174,412,286,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000971.jpg 221,49,243,71,0 210,202,224,216,0 35,19,57,41,2 28,278,70,320,6 221,197,277,253,9 220,116,262,158,5 236,252,320,336,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000972.jpg 293,183,315,205,2 310,25,324,39,4 32,28,88,84,6 198,243,254,299,0 104,200,132,228,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000973.jpg 90,76,112,98,9 60,3,102,45,4 1,246,43,288,0 113,335,169,391,5 1,303,85,387,6 187,21,271,105,3 10,133,122,245,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000974.jpg 83,335,97,349,8 279,135,301,157,2 264,279,278,293,8 16,284,44,312,9 250,166,292,208,8 288,162,372,246,5 241,265,353,377,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000975.jpg 188,346,202,360,2 262,337,318,393,2 64,48,92,76,4 203,257,259,313,1 52,282,108,338,8 174,358,230,414,4 342,26,384,68,9 113,172,197,256,8 323,242,407,326,3 236,113,348,225,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000976.jpg 77,298,99,320,5 150,236,172,258,7 106,314,162,370,1 349,323,391,365,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000977.jpg 59,27,73,41,9 119,258,133,272,6 46,227,158,339,3 9,90,121,202,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000978.jpg 105,286,127,308,0 20,399,34,413,6 335,138,391,194,2 268,152,296,180,5 118,345,146,373,8 183,156,267,240,9 72,115,156,199,2 32,197,116,281,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000979.jpg 40,213,82,255,3 348,170,376,198,9 108,333,150,375,7 137,96,179,138,4 188,183,272,267,2 273,193,385,305,1 33,256,117,340,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000980.jpg 16,132,30,146,9 49,297,105,353,1 141,261,183,303,3 263,254,319,310,6 204,273,232,301,2 360,197,388,225,9 358,125,386,153,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000981.jpg 180,23,194,37,4 17,63,45,91,2 302,256,344,298,6 289,39,373,123,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000982.jpg 188,83,244,139,7 338,167,394,223,4 239,0,267,28,0 177,287,289,399,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000983.jpg 22,0,44,22,7 263,336,319,392,3 22,184,64,226,2 315,227,357,269,6 61,25,89,53,0 162,229,274,341,5 41,242,153,354,9 190,120,274,204,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000984.jpg 128,123,240,235,6 130,313,214,397,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000985.jpg 27,376,41,390,0 372,309,400,337,6 303,135,415,247,7 147,4,231,88,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000986.jpg 199,348,221,370,8 121,65,163,107,9 130,110,172,152,9 42,304,70,332,9 12,28,124,140,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000987.jpg 297,187,311,201,5 94,138,116,160,5 166,47,250,131,1 9,13,121,125,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000988.jpg 382,84,404,106,1 196,51,224,79,5 261,230,289,258,6 164,104,192,132,2 39,139,67,167,8 211,248,253,290,1 77,299,133,355,0 311,216,395,300,1 184,296,268,380,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000989.jpg 283,65,305,87,8 173,255,195,277,9 67,275,179,387,2 191,285,303,397,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000990.jpg 232,87,288,143,5 81,230,165,314,2 81,111,193,223,4 271,199,383,311,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000991.jpg 353,197,375,219,2 227,155,249,177,5 252,33,266,47,0 223,280,335,392,1 259,184,343,268,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000992.jpg 314,211,336,233,5 265,34,321,90,3 4,152,60,208,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000993.jpg 94,138,108,152,7 378,25,400,47,7 324,208,352,236,3 345,118,373,146,4 215,281,243,309,3 298,253,410,365,9 229,42,341,154,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000994.jpg 49,362,71,384,6 340,283,382,325,3 42,187,70,215,0 162,372,204,414,6 291,170,319,198,0 169,13,253,97,4 160,187,244,271,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000995.jpg 14,1,126,113,0 34,171,118,255,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000996.jpg 366,340,380,354,9 199,191,213,205,5 335,2,357,24,5 112,189,140,217,3 153,91,181,119,8 149,262,233,346,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000997.jpg 49,32,71,54,3 130,156,186,212,9 59,312,101,354,5 129,278,185,334,6 369,382,397,410,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000998.jpg 344,49,400,105,8 209,222,265,278,9 79,137,135,193,7 322,287,406,371,7 49,11,133,95,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/000999.jpg 162,145,184,167,2 258,26,370,138,5 176,149,288,261,2 29,310,113,394,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_train/001000.jpg 243,105,265,127,4 143,155,165,177,6 19,239,75,295,4 136,245,164,273,6 35,10,77,52,1 333,169,375,211,5 239,357,281,399,7 230,206,342,318,4 87,18,199,130,9 294,45,406,157,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000001.jpg 227,246,241,260,9 401,50,415,64,7 20,79,76,135,3 138,112,194,168,8 244,31,328,115,9 273,257,385,369,1 121,172,205,256,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000002.jpg 26,90,82,146,5 332,100,360,128,8 165,265,221,321,1 18,183,130,295,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000003.jpg 121,137,143,159,3 314,313,342,341,1 60,184,102,226,4 363,178,391,206,9 125,264,181,320,9 350,74,392,116,4 169,182,225,238,2 58,29,142,113,6 198,44,282,128,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000004.jpg 24,166,46,188,0 190,165,204,179,9 98,255,154,311,0 325,195,367,237,5 216,328,300,412,9 98,49,182,133,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000005.jpg 107,181,149,223,1 204,82,316,194,1 109,256,193,340,9 193,299,305,411,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000006.jpg 204,318,226,340,8 107,205,121,219,5 129,297,143,311,9 297,262,353,318,7 22,173,78,229,1 87,10,171,94,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000007.jpg 213,324,297,408,7 36,307,120,391,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000008.jpg 197,259,239,301,7 72,183,184,295,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000009.jpg 295,183,351,239,1 315,105,371,161,9 245,62,273,90,3 171,356,199,384,9 366,59,408,101,8 85,37,169,121,1 169,250,281,362,5 92,149,176,233,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000010.jpg 344,95,358,109,0 195,334,217,356,8 251,135,307,191,6 146,296,188,338,5 241,319,297,375,4 280,285,308,313,3 180,15,236,71,8 107,115,219,227,3 43,42,127,126,3 300,189,412,301,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000011.jpg 94,320,108,334,2 308,234,322,248,2 38,255,52,269,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000012.jpg 217,44,239,66,0 249,245,305,301,6 31,13,87,69,6 321,264,349,292,9 352,193,380,221,7 109,294,165,350,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000013.jpg 161,190,183,212,9 163,290,177,304,6 213,363,235,385,6 258,246,300,288,1 334,345,390,401,8 85,198,127,240,5 53,371,95,413,8 20,201,62,243,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000014.jpg 333,359,355,381,8 15,117,37,139,1 198,331,240,373,5 29,358,85,414,1 361,294,403,336,3 285,54,369,138,1 114,45,198,129,6 216,179,328,291,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000015.jpg 59,304,87,332,5 299,13,327,41,1 311,81,367,137,1 171,9,213,51,9 232,133,274,175,8 159,199,215,255,5 289,241,401,353,8 104,62,188,146,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000016.jpg 219,244,233,258,0 232,356,288,412,0 19,66,61,108,8 40,117,96,173,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000017.jpg 19,278,61,320,2 162,35,218,91,3 383,117,411,145,9 259,79,371,191,6 53,122,165,234,1 125,275,237,387,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000018.jpg 58,368,72,382,7 215,44,243,72,9 309,90,337,118,6 53,76,109,132,6 86,330,142,386,5 310,279,366,335,1 82,152,194,264,4 222,148,334,260,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000019.jpg 261,70,289,98,1 125,186,181,242,1 315,3,399,87,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000020.jpg 293,32,315,54,8 119,122,133,136,0 138,81,152,95,7 122,264,150,292,7 349,270,391,312,3 224,299,308,383,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000021.jpg 43,39,57,53,7 347,392,361,406,2 224,329,252,357,3 312,93,340,121,3 359,207,401,249,3 109,369,137,397,6 7,167,63,223,7 143,332,185,374,2 252,256,336,340,5 87,177,171,261,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000022.jpg 244,41,286,83,3 113,230,197,314,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000023.jpg 159,40,173,54,3 201,262,215,276,9 384,158,406,180,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000024.jpg 22,55,44,77,8 205,0,227,22,3 318,272,340,294,3 161,321,203,363,9 118,118,146,146,0 268,191,324,247,1 246,284,274,312,5 141,100,253,212,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000025.jpg 156,148,212,204,5 304,241,332,269,3 218,238,274,294,2 292,274,334,316,2 120,223,176,279,2 38,112,66,140,6 252,48,364,160,8 126,294,238,406,4 86,30,198,142,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000026.jpg 282,244,296,258,8 147,278,161,292,4 271,82,293,104,0 173,152,285,264,3 296,213,408,325,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000027.jpg 275,140,289,154,1 216,260,238,282,9 82,319,104,341,1 258,273,286,301,3 186,92,242,148,5 183,5,239,61,6 129,261,185,317,7 78,123,162,207,7 278,194,362,278,5 263,19,347,103,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000028.jpg 57,43,71,57,5 101,66,115,80,7 349,245,363,259,8 368,285,396,313,0 335,149,391,205,9 149,216,191,258,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000029.jpg 374,55,396,77,2 51,116,65,130,6 120,120,134,134,8 213,157,241,185,0 28,224,70,266,8 187,30,229,72,2 187,331,229,373,1 186,201,298,313,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000030.jpg 320,74,342,96,1 25,384,39,398,7 144,185,228,269,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000031.jpg 42,197,64,219,0 104,148,216,260,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000032.jpg 164,173,192,201,1 121,19,149,47,2 287,139,315,167,8 310,87,394,171,5 5,205,117,317,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000033.jpg 317,248,331,262,1 299,339,321,361,5 284,229,306,251,5 332,16,360,44,9 155,3,211,59,7 292,358,348,414,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000034.jpg 208,260,230,282,0 12,2,26,16,0 25,318,39,332,7 113,38,169,94,2 84,165,126,207,4 306,215,390,299,9 86,282,198,394,3 209,0,321,112,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000035.jpg 25,265,47,287,9 122,52,178,108,5 332,226,388,282,0 348,46,404,102,5 71,51,99,79,0 4,148,60,204,4 193,293,277,377,9 50,175,162,287,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000036.jpg 6,343,62,399,7 373,158,401,186,4 351,310,407,366,5 276,20,318,62,1 146,215,174,243,6 43,204,127,288,1 179,292,291,404,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000037.jpg 54,265,68,279,1 129,40,151,62,8 192,319,220,347,5 292,312,320,340,1 333,289,389,345,1 112,317,168,373,0 301,124,357,180,8 55,164,139,248,2 311,41,395,125,2 182,110,294,222,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000038.jpg 48,91,62,105,8 26,106,54,134,1 215,305,243,333,6 330,12,386,68,3 338,104,394,160,0 73,108,185,220,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000039.jpg 382,234,396,248,4 96,82,152,138,6 254,2,310,58,3 296,182,352,238,4 111,299,223,411,6 28,193,140,305,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000040.jpg 360,4,382,26,1 374,33,396,55,4 38,326,52,340,0 304,320,346,362,8 345,202,387,244,8 108,247,164,303,7 358,70,386,98,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000041.jpg 70,391,84,405,7 217,382,239,404,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000042.jpg 266,89,288,111,8 286,283,300,297,9 121,323,163,365,4 271,33,313,75,3 373,212,415,254,7 35,266,119,350,2 13,154,125,266,9 266,243,378,355,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000043.jpg 226,142,240,156,2 128,92,142,106,4 293,122,377,206,1 164,205,276,317,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000044.jpg 135,208,149,222,3 161,377,175,391,3 388,308,410,330,4 26,150,68,192,6 110,67,152,109,0 330,318,372,360,8 144,121,200,177,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000045.jpg 96,114,152,170,8 210,233,238,261,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000046.jpg 274,370,288,384,1 108,53,122,67,3 42,45,64,67,1 121,85,177,141,7 213,43,325,155,7 206,293,290,377,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000047.jpg 26,334,48,356,3 277,229,305,257,7 105,24,147,66,9 288,280,344,336,2 204,302,232,330,0 178,80,206,108,2 55,64,97,106,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000048.jpg 318,57,340,79,2 312,232,326,246,1 273,128,287,142,9 8,282,50,324,2 307,308,335,336,2 113,9,197,93,5 272,129,384,241,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000049.jpg 113,14,127,28,2 351,388,373,410,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000050.jpg 286,272,308,294,6 238,165,252,179,2 365,57,407,99,3 118,201,230,313,2 125,30,237,142,9 36,315,120,399,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000051.jpg 295,384,317,406,2 21,227,43,249,1 10,298,52,340,3 69,314,111,356,6 170,102,212,144,0 117,186,145,214,7 37,122,79,164,8 90,73,174,157,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000052.jpg 198,77,220,99,8 143,212,199,268,1 26,1,54,29,4 293,71,321,99,0 325,216,353,244,1 156,90,198,132,0 276,209,318,251,5 34,52,146,164,1 170,283,282,395,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000053.jpg 59,238,73,252,1 370,77,392,99,7 96,157,118,179,1 27,170,83,226,9 83,22,167,106,9 278,261,362,345,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000054.jpg 277,160,291,174,3 288,24,310,46,2 109,187,137,215,2 107,226,219,338,0 13,115,97,199,2 303,300,415,412,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000055.jpg 307,315,321,329,8 339,195,381,237,5 156,255,198,297,1 9,55,51,97,7 181,327,209,355,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000056.jpg 160,351,182,373,2 386,390,400,404,3 373,235,387,249,4 196,64,238,106,9 216,22,258,64,1 139,384,167,412,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000057.jpg 184,101,198,115,5 156,293,240,377,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000058.jpg 209,213,231,235,6 164,221,186,243,9 77,29,99,51,7 19,313,75,369,9 337,171,393,227,5 226,23,338,135,2 108,70,220,182,8 282,239,394,351,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000059.jpg 99,334,121,356,5 196,54,238,96,5 335,369,377,411,6 224,192,308,276,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000060.jpg 375,380,397,402,3 87,282,143,338,8 73,211,129,267,5 72,79,156,163,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000061.jpg 337,295,379,337,5 40,41,96,97,4 8,214,50,256,1 144,128,228,212,0 99,247,211,359,0 100,16,212,128,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000062.jpg 127,126,183,182,8 234,133,290,189,0 117,255,173,311,7 297,9,409,121,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000063.jpg 112,207,126,221,4 275,191,297,213,8 226,193,240,207,2 270,16,298,44,6 292,107,404,219,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000064.jpg 264,5,286,27,5 282,297,296,311,2 19,331,41,353,9 116,271,144,299,3 78,186,120,228,6 69,297,111,339,1 324,323,352,351,1 126,147,238,259,0 279,163,363,247,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000065.jpg 103,55,145,97,9 261,237,317,293,4 323,54,351,82,2 202,23,258,79,5 278,140,320,182,9 96,188,152,244,8 160,95,244,179,7 4,14,88,98,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000066.jpg 333,292,355,314,7 24,13,38,27,6 14,237,42,265,7 278,11,334,67,6 47,284,75,312,9 72,81,128,137,2 6,152,48,194,9 205,60,317,172,2 195,224,279,308,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000067.jpg 167,100,181,114,0 351,120,365,134,4 172,286,214,328,3 340,228,396,284,4 337,336,379,378,7 224,71,308,155,1 93,113,205,225,5 6,2,118,114,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000068.jpg 266,302,294,330,7 328,171,356,199,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000069.jpg 64,358,86,380,7 175,185,189,199,5 328,271,384,327,7 178,236,206,264,3 285,55,313,83,4 52,118,164,230,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000070.jpg 131,379,153,401,0 310,163,332,185,9 63,369,85,391,3 126,352,154,380,2 298,81,326,109,3 322,24,378,80,3 37,90,79,132,4 356,373,398,415,5 188,34,230,76,4 159,172,243,256,8 134,76,218,160,8 261,282,345,366,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000071.jpg 71,395,85,409,8 348,66,370,88,8 320,160,362,202,9 218,361,246,389,3 252,275,364,387,2 153,236,237,320,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000072.jpg 222,357,278,413,4 346,174,374,202,2 201,93,229,121,8 148,159,260,271,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000073.jpg 137,94,159,116,2 222,287,244,309,4 339,278,381,320,9 62,291,104,333,4 81,28,137,84,3 9,182,51,224,7 194,4,250,60,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000074.jpg 99,93,121,115,2 31,332,87,388,6 179,208,207,236,5 37,153,93,209,5 382,258,410,286,8 83,284,195,396,4 250,72,362,184,6 121,81,233,193,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000075.jpg 359,186,401,228,5 159,302,215,358,0 212,241,254,283,6 122,281,150,309,4 120,109,148,137,3 112,345,140,373,5 142,112,226,196,7 228,5,312,89,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000076.jpg 77,372,91,386,0 36,6,50,20,9 324,314,408,398,7 17,98,129,210,5 159,143,271,255,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000077.jpg 151,134,173,156,0 204,33,218,47,7 11,69,123,181,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000078.jpg 122,262,144,284,1 157,144,185,172,7 319,346,375,402,4 68,312,152,396,5 137,222,249,334,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000079.jpg 36,197,50,211,0 391,144,413,166,7 143,68,185,110,1 248,239,360,351,1 286,60,370,144,1 106,134,190,218,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000080.jpg 310,138,332,160,3 165,119,249,203,6 78,207,162,291,2 264,281,348,365,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000081.jpg 2,13,24,35,3 153,264,175,286,6 207,247,319,359,3 259,38,343,122,6 174,107,258,191,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000082.jpg 182,203,204,225,0 67,190,81,204,7 361,161,375,175,6 214,277,256,319,8 114,286,170,342,7 133,6,175,48,7 214,375,242,403,8 6,286,118,398,0 284,284,396,396,9 238,70,322,154,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000083.jpg 159,171,201,213,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000084.jpg 153,128,209,184,9 178,52,220,94,7 354,277,410,333,5 198,286,282,370,1 87,319,171,403,1 230,99,342,211,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000085.jpg 299,77,313,91,2 66,178,80,192,7 173,135,187,149,3 114,364,142,392,1 244,245,272,273,4 148,271,204,327,0 339,102,381,144,7 302,185,386,269,1 98,125,182,209,1 15,246,127,358,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000086.jpg 43,5,57,19,6 345,63,373,91,8 299,231,411,343,7 8,38,120,150,0 230,117,314,201,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000087.jpg 2,305,16,319,2 174,299,216,341,6 60,189,102,231,9 249,243,277,271,6 298,219,382,303,5 319,35,403,119,2 177,14,261,98,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000088.jpg 322,319,344,341,4 122,221,164,263,8 365,356,393,384,3 188,224,272,308,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000089.jpg 64,252,78,266,1 184,174,198,188,3 69,22,125,78,8 155,45,267,157,6 260,261,344,345,5 15,288,99,372,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000090.jpg 252,299,294,341,4 340,17,368,45,2 128,36,170,78,9 339,88,367,116,9 65,16,121,72,8 217,73,301,157,4 3,284,115,396,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000091.jpg 347,354,389,396,8 29,98,57,126,0 307,227,349,269,2 221,174,305,258,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000092.jpg 302,258,324,280,1 322,66,364,108,7 176,97,204,125,1 2,2,86,86,0 49,106,161,218,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000093.jpg 73,85,87,99,7 164,213,178,227,3 193,298,221,326,8 28,228,70,270,5 219,381,247,409,5 102,87,144,129,9 144,37,186,79,1 260,52,302,94,5 211,188,295,272,3 78,184,190,296,1 230,269,342,381,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000094.jpg 368,205,396,233,3 150,41,206,97,0 204,358,246,400,9 59,170,171,282,8 260,139,344,223,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000095.jpg 66,107,94,135,9 152,289,180,317,7 214,78,270,134,1 54,166,110,222,0 357,97,385,125,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000096.jpg 24,167,38,181,2 308,15,322,29,6 50,257,106,313,8 234,94,290,150,8 185,275,241,331,2 68,349,96,377,5 137,219,179,261,6 110,177,152,219,7 97,38,209,150,6 292,258,404,370,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000097.jpg 317,130,373,186,2 2,129,44,171,6 184,183,240,239,2 303,261,345,303,8 116,168,172,224,6 118,263,174,319,1 70,22,182,134,5 247,299,331,383,1 13,319,97,403,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000098.jpg 145,335,159,349,9 99,359,155,415,7 138,46,180,88,8 86,150,128,192,7 165,287,221,343,1 234,105,262,133,8 300,92,412,204,8 34,226,146,338,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000099.jpg 251,154,265,168,5 357,112,379,134,4 147,256,203,312,4 208,149,250,191,1 354,238,396,280,3 81,187,137,243,4 90,321,132,363,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000100.jpg 176,155,232,211,0 30,176,58,204,3 215,3,271,59,6 360,217,388,245,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000101.jpg 398,162,412,176,1 120,360,134,374,9 287,153,301,167,1 71,93,127,149,0 141,250,183,292,1 5,106,61,162,9 167,24,209,66,6 333,68,361,96,3 12,266,96,350,6 290,123,402,235,9 210,274,294,358,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000102.jpg 388,103,410,125,8 108,276,136,304,2 373,321,401,349,4 359,31,401,73,7 51,211,79,239,0 161,234,273,346,0 21,315,105,399,0 286,290,370,374,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000103.jpg 376,112,398,134,8 268,347,290,369,7 208,246,230,268,3 299,162,355,218,7 378,207,406,235,8 64,61,176,173,0 101,276,213,388,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000104.jpg 47,88,89,130,5 197,116,225,144,9 246,166,288,208,3 51,274,93,316,1 206,22,262,78,3 277,35,333,91,9 232,202,344,314,9 109,17,193,101,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000105.jpg 218,119,240,141,6 34,132,48,146,2 113,290,141,318,7 108,32,150,74,1 213,157,269,213,8 293,292,335,334,4 251,7,335,91,0 58,97,170,209,0 35,297,119,381,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000106.jpg 46,128,60,142,7 391,306,413,328,9 146,220,188,262,0 202,131,286,215,8 106,299,218,411,4 92,136,176,220,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000107.jpg 309,72,331,94,5 74,271,88,285,4 21,203,63,245,8 304,229,332,257,3 159,186,215,242,8 135,32,177,74,1 212,329,240,357,8 57,325,113,381,9 61,75,173,187,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000108.jpg 260,17,274,31,6 299,63,355,119,0 24,253,80,309,0 345,338,373,366,9 280,190,322,232,7 208,321,250,363,1 102,126,214,238,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000109.jpg 335,283,349,297,1 278,208,320,250,7 224,255,252,283,1 292,132,348,188,1 33,163,89,219,0 220,149,276,205,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000110.jpg 12,121,26,135,6 23,181,45,203,7 100,63,114,77,3 335,198,363,226,3 115,165,157,207,9 212,145,296,229,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000111.jpg 17,113,39,135,1 354,215,396,257,2 96,351,152,407,4 61,252,89,280,8 351,17,393,59,1 150,132,234,216,7 221,7,305,91,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000112.jpg 127,209,141,223,2 8,296,50,338,9 69,266,111,308,1 328,239,356,267,6 334,77,376,119,5 220,19,304,103,1 172,326,256,410,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000113.jpg 368,324,390,346,7 267,362,289,384,5 239,36,295,92,9 81,19,165,103,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000114.jpg 120,271,134,285,5 199,161,221,183,8 115,195,157,237,1 54,350,82,378,5 22,0,134,112,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000115.jpg 140,178,154,192,1 146,314,174,342,6 286,254,342,310,7 257,335,313,391,7 28,17,70,59,6 31,101,59,129,3 208,243,250,285,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000116.jpg 26,199,110,283,4 68,22,180,134,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000117.jpg 240,96,262,118,0 42,189,56,203,6 75,318,131,374,9 289,178,401,290,8 143,303,255,415,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000118.jpg 299,16,313,30,7 160,49,174,63,4 76,214,104,242,0 15,112,71,168,9 258,348,300,390,5 150,106,262,218,9 120,276,232,388,6 275,66,387,178,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000119.jpg 78,328,120,370,7 59,5,101,47,7 113,370,141,398,8 207,255,235,283,5 142,152,198,208,9 198,170,282,254,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000120.jpg 228,313,242,327,2 186,212,228,254,0 122,5,178,61,9 301,252,343,294,1 45,354,101,410,3 230,148,258,176,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000121.jpg 148,252,190,294,7 152,66,208,122,6 125,194,181,250,6 6,171,90,255,9 207,155,291,239,6 72,326,156,410,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000122.jpg 23,266,45,288,3 173,232,201,260,0 222,61,264,103,8 44,318,100,374,7 306,195,348,237,4 221,273,305,357,2 97,246,181,330,1 74,126,158,210,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000123.jpg 38,287,52,301,2 286,6,314,34,1 107,58,149,100,4 118,91,230,203,8 296,170,380,254,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000124.jpg 116,331,144,359,4 214,308,242,336,0 115,72,199,156,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000125.jpg 193,182,221,210,9 60,177,102,219,2 363,363,405,405,0 330,217,372,259,1 308,121,336,149,9 245,260,301,316,8 113,72,225,184,9 65,267,177,379,4 259,1,371,113,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000126.jpg 232,203,288,259,2 28,167,56,195,9 115,19,157,61,6 334,278,376,320,7 23,16,65,58,0 319,320,403,404,3 2,238,86,322,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000127.jpg 1,358,15,372,9 141,256,155,270,5 97,150,153,206,1 44,178,100,234,9 116,212,144,240,3 281,99,323,141,2 289,321,317,349,7 365,5,407,47,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000128.jpg 75,273,187,385,0 238,146,350,258,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000129.jpg 304,33,326,55,4 329,52,413,136,0 69,17,153,101,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000130.jpg 4,348,18,362,8 356,163,378,185,3 106,316,190,400,5 30,97,114,181,7 248,100,332,184,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000131.jpg 28,109,42,123,6 291,2,333,44,7 370,281,398,309,8 117,94,159,136,7 232,204,316,288,5 256,73,368,185,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000132.jpg 22,220,78,276,2 72,264,100,292,4 169,152,281,264,6 180,266,292,378,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000133.jpg 170,201,184,215,9 369,213,391,235,1 277,149,291,163,3 34,22,90,78,9 310,45,366,101,3 149,64,233,148,1 254,97,366,209,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000134.jpg 71,99,85,113,9 95,134,117,156,2 328,100,356,128,0 154,19,210,75,4 139,124,167,152,4 270,323,312,365,6 186,176,270,260,8 275,171,387,283,9 37,328,121,412,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000135.jpg 359,23,373,37,6 141,107,155,121,1 11,127,25,141,4 342,220,370,248,5 220,60,332,172,6 112,307,196,391,7 238,300,350,412,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000136.jpg 214,315,242,343,4 41,207,97,263,7 329,367,357,395,4 326,187,368,229,4 323,144,365,186,6 307,13,335,41,3 305,253,389,337,1 31,6,143,118,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000137.jpg 122,369,164,411,3 326,110,382,166,0 70,296,126,352,4 326,245,354,273,1 156,278,240,362,3 133,135,217,219,9 11,143,123,255,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000138.jpg 9,183,37,211,9 51,200,79,228,4 153,20,237,104,9 35,279,147,391,9 196,106,308,218,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000139.jpg 273,195,295,217,1 338,47,394,103,0 370,181,412,223,8 189,24,217,52,3 207,355,235,383,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000140.jpg 323,199,337,213,0 150,108,164,122,4 86,292,108,314,1 267,227,323,283,3 140,359,168,387,6 111,3,167,59,7 367,340,409,382,7 194,147,250,203,6 123,292,151,320,8 274,74,358,158,7 52,78,164,190,9 214,305,298,389,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000141.jpg 186,365,208,387,1 23,199,65,241,6 119,67,147,95,7 263,32,319,88,2 208,58,250,100,7 169,212,281,324,0 330,94,414,178,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000142.jpg 20,119,34,133,1 192,73,214,95,0 395,3,409,17,6 220,141,248,169,1 347,57,403,113,7 189,351,245,407,1 55,181,167,293,7 228,210,340,322,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000143.jpg 336,384,358,406,8 90,397,104,411,7 339,270,381,312,9 84,95,112,123,3 270,70,312,112,4 187,215,299,327,8 82,128,194,240,1 286,133,398,245,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000144.jpg 111,87,125,101,1 232,234,254,256,3 27,304,111,388,7 118,265,202,349,0 272,108,356,192,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000145.jpg 376,106,390,120,2 114,18,128,32,3 300,295,356,351,8 274,339,302,367,8 297,194,339,236,6 384,164,412,192,7 124,202,180,258,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000146.jpg 250,268,292,310,1 281,322,323,364,1 89,26,131,68,4 278,75,334,131,1 138,194,250,306,7 135,107,219,191,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000147.jpg 146,273,168,295,0 354,65,368,79,7 93,24,107,38,4 33,77,61,105,7 229,76,341,188,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000148.jpg 108,145,130,167,2 143,276,165,298,2 127,184,183,240,0 197,221,253,277,1 142,126,184,168,1 269,37,325,93,1 19,255,131,367,9 27,7,111,91,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000149.jpg 33,75,55,97,9 268,197,282,211,6 284,72,306,94,4 375,200,403,228,1 69,174,125,230,3 20,9,48,37,3 176,268,288,380,7 96,41,208,153,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000150.jpg 27,9,41,23,7 269,136,297,164,6 2,43,86,127,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000151.jpg 280,219,336,275,4 170,178,226,234,8 146,273,202,329,0 49,253,91,295,1 198,90,226,118,1 15,369,57,411,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000152.jpg 306,193,320,207,2 384,204,398,218,4 75,234,131,290,0 197,387,225,415,5 203,44,287,128,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000153.jpg 204,102,218,116,9 234,17,256,39,4 343,356,365,378,2 186,196,228,238,8 54,229,110,285,3 192,39,304,151,6 247,155,331,239,2 199,303,283,387,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000154.jpg 16,29,38,51,1 368,56,390,78,2 317,28,345,56,8 287,71,343,127,4 273,3,315,45,3 141,176,253,288,4 52,284,164,396,2 42,96,126,180,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000155.jpg 273,30,287,44,8 199,23,213,37,6 301,333,315,347,2 183,347,225,389,2 55,80,111,136,3 116,37,200,121,1 301,82,385,166,0 63,301,147,385,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000156.jpg 255,305,269,319,2 73,216,95,238,6 67,357,95,385,2 87,245,115,273,7 120,300,232,412,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000157.jpg 371,339,385,353,1 135,216,157,238,3 159,259,187,287,5 68,230,110,272,1 150,307,234,391,0 13,42,97,126,2 103,42,187,126,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000158.jpg 290,178,304,192,6 24,162,38,176,9 91,249,133,291,1 84,15,140,71,3 6,331,62,387,0 147,208,231,292,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000159.jpg 319,165,341,187,6 356,105,384,133,2 8,34,64,90,3 148,15,190,57,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000160.jpg 351,116,373,138,9 268,337,282,351,9 219,106,241,128,8 328,153,356,181,7 282,275,338,331,0 135,128,191,184,5 2,309,30,337,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000161.jpg 143,391,165,413,3 226,23,240,37,0 150,261,192,303,0 168,204,196,232,1 105,203,133,231,1 335,275,391,331,0 200,101,256,157,5 23,144,79,200,6 82,98,194,210,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000162.jpg 314,205,328,219,3 36,11,50,25,3 177,165,219,207,2 132,351,174,393,2 110,292,138,320,5 375,314,403,342,2 319,345,361,387,5 66,42,178,154,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000163.jpg 214,14,270,70,1 343,43,399,99,7 209,350,265,406,5 365,147,407,189,2 237,128,321,212,1 75,148,159,232,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000164.jpg 293,213,307,227,2 287,40,315,68,4 317,300,359,342,7 8,247,50,289,5 349,177,405,233,4 115,293,143,321,3 146,109,230,193,8 168,279,252,363,3 223,177,335,289,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000165.jpg 155,139,211,195,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000166.jpg 319,3,341,25,3 3,127,25,149,7 224,141,238,155,7 167,314,209,356,8 244,48,272,76,8 214,188,256,230,8 79,132,107,160,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000167.jpg 61,340,75,354,0 150,358,172,380,4 131,12,153,34,9 260,302,302,344,2 324,24,380,80,0 172,177,256,261,1 273,78,385,190,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000168.jpg 292,25,320,53,0 23,27,65,69,7 116,59,144,87,8 157,193,185,221,2 209,117,293,201,6 38,201,122,285,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000169.jpg 360,382,374,396,0 216,177,272,233,8 77,266,133,322,6 6,337,62,393,9 120,150,204,234,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000170.jpg 67,361,81,375,1 46,383,60,397,5 12,290,34,312,0 372,283,414,325,7 62,73,104,115,6 194,17,306,129,9 32,144,116,228,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000171.jpg 38,320,52,334,3 138,347,152,361,0 10,233,66,289,4 117,151,159,193,1 244,265,286,307,9 334,379,362,407,8 265,23,377,135,8 4,102,116,214,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000172.jpg 178,334,206,362,7 277,268,333,324,0 2,27,114,139,6 162,189,246,273,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000173.jpg 401,37,415,51,7 228,69,242,83,9 284,191,298,205,8 155,247,197,289,9 312,140,340,168,6 172,204,214,246,6 65,244,93,272,5 72,47,184,159,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000174.jpg 367,108,395,136,8 154,10,196,52,5 258,298,286,326,7 139,288,181,330,7 46,76,88,118,1 276,127,388,239,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000175.jpg 196,383,210,397,8 44,29,66,51,3 295,1,323,29,2 333,319,361,347,6 316,144,358,186,7 250,332,306,388,9 135,109,177,151,0 52,192,136,276,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000176.jpg 360,62,374,76,0 224,301,246,323,5 330,312,352,334,5 157,140,199,182,9 82,254,124,296,1 344,378,372,406,9 81,213,123,255,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000177.jpg 85,22,99,36,6 113,171,135,193,3 20,263,42,285,0 316,193,372,249,0 82,304,124,346,2 4,133,46,175,6 55,38,139,122,8 229,284,341,396,4 315,95,399,179,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000178.jpg 133,154,147,168,0 381,260,403,282,1 180,179,194,193,2 251,129,279,157,2 119,42,175,98,2 149,293,191,335,6 255,10,297,52,0 78,291,134,347,1 288,161,372,245,5 7,45,91,129,8 267,282,351,366,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000179.jpg 218,207,232,221,0 69,336,97,364,1 68,9,110,51,6 11,93,123,205,8 138,6,250,118,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000180.jpg 126,122,140,136,9 311,292,333,314,8 266,170,378,282,8 9,93,93,177,0 132,43,216,127,3\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000181.jpg 116,64,138,86,5 104,309,126,331,2 123,360,145,382,9 141,93,169,121,2 65,76,107,118,5 375,291,403,319,9 136,245,248,357,3 128,127,212,211,7 290,221,374,305,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000182.jpg 350,360,372,382,6 199,273,241,315,4 88,226,116,254,7 104,65,132,93,3 266,96,322,152,6 236,135,264,163,6 137,315,179,357,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000183.jpg 74,24,96,46,8 9,360,23,374,3 125,8,181,64,5 210,114,266,170,4 171,73,213,115,3 309,170,337,198,6 283,56,325,98,1 5,154,61,210,2 97,286,209,398,1 255,200,367,312,5 300,307,384,391,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000184.jpg 273,165,315,207,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000185.jpg 205,235,261,291,8 310,71,366,127,2 74,263,158,347,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000186.jpg 341,312,355,326,8 298,64,354,120,2 150,219,192,261,5 214,191,270,247,0 290,123,374,207,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000187.jpg 305,16,327,38,1 177,181,191,195,1 173,52,187,66,9 208,170,320,282,3 48,225,132,309,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000188.jpg 239,333,253,347,4 160,281,244,365,0 25,108,137,220,8 299,156,383,240,0\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000189.jpg 354,57,368,71,7 306,0,362,56,0 87,164,115,192,3 206,373,234,401,2\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000190.jpg 46,209,60,223,3 380,194,408,222,6 349,221,405,277,0 204,143,232,171,5 132,278,244,390,9 150,26,262,138,2 28,81,112,165,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000191.jpg 367,178,395,206,0 95,80,151,136,1 142,311,198,367,2 39,6,67,34,9 275,183,303,211,6 245,266,329,350,9 121,208,205,292,3 24,204,108,288,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000192.jpg 385,81,407,103,2 145,329,167,351,9 61,288,103,330,7 228,368,256,396,0 131,79,187,135,2 230,207,286,263,8 169,342,225,398,8\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000193.jpg 232,258,288,314,9 36,239,92,295,3 296,233,324,261,0 363,370,405,412,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000194.jpg 120,331,134,345,7 207,20,221,34,6 175,251,189,265,3 232,252,288,308,2 6,12,62,68,2 176,174,232,230,9 47,234,103,290,7 188,313,244,369,9\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000195.jpg 6,393,20,407,7 101,41,157,97,6\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000196.jpg 116,148,130,162,8 339,179,395,235,1 52,254,108,310,1\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000197.jpg 337,199,359,221,4 262,140,276,154,9 55,107,69,121,4 177,224,233,280,4 301,274,329,302,2 65,330,93,358,6 227,168,283,224,4 9,151,51,193,0 291,65,403,177,1 142,316,226,400,5\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000198.jpg 278,367,292,381,3 343,349,357,363,6 298,138,320,160,8 245,254,273,282,9 153,171,237,255,4\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000199.jpg 358,313,372,327,4 159,333,201,375,1 135,232,191,288,4 128,81,156,109,7\n",
            "=> /content/TensorFlow-2.x-YOLOv3/mnist/mnist_test/000200.jpg 138,61,166,89,4 62,23,104,65,8 270,197,298,225,4 278,246,320,288,0 151,209,179,237,1 257,89,369,201,0 117,229,229,341,3 36,84,148,196,1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DDZpw-RLMX8h",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OGAKWzdRN_l3",
        "colab_type": "text"
      },
      "source": [
        "## Model YOLO"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5RIzKhSfNV8W",
        "colab_type": "text"
      },
      "source": [
        "### Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9Elpp7VnLgTs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from __future__ import absolute_import, division, print_function, unicode_literals\n",
        "from tensorflow.keras import datasets, layers, models\n",
        "from tensorflow.keras import backend as K\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.layers import Input\n",
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.layers import Layer, InputSpec, Concatenate, LeakyReLU\n",
        "from tensorflow.keras.layers import Conv2D, Input, LeakyReLU, ZeroPadding2D, BatchNormalization, MaxPool2D\n",
        "from tensorflow.keras.regularizers import l2\n",
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import shutil\n",
        "import os"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jgdxHNTlLs_b",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "STRIDES = [8, 16, 32]\n",
        "ANCHORS = [[[10,  13], [16,   30], [33,   23]],\n",
        "               [[30,  61], [62,   45], [59,  119]],\n",
        "               [[116, 90], [156, 198], [373, 326]]]\n",
        "TRAIN_WARMUP_EPOCHS = 1\n",
        "TRAIN_EPOCHS = 10\n",
        "TRAIN_LOGDIR = './yolo_checkpoints'\n",
        "TRAIN_YOLO_TINY = False\n",
        "TRAIN_CLASSES               = \"mnist/mnist.names\"\n",
        "TRAIN_ANNOT_PATH            = \"mnist/mnist_train.txt\"\n",
        "TRAIN_LOGDIR                = \"log\"\n",
        "TRAIN_CHECKPOINTS_FOLDER    = \"checkpoints\"\n",
        "TRAIN_BATCH_SIZE            = 8\n",
        "TRAIN_INPUT_SIZE            = 416\n",
        "TRAIN_DATA_AUG              = True\n",
        "TRAIN_LR_INIT               = 1e-4\n",
        "TRAIN_LR_END                = 1e-6\n",
        "TRAIN_WARMUP_EPOCHS         = 2\n",
        "TRAIN_EPOCHS                = 100\n",
        "TEST_ANNOT_PATH             = \"mnist/mnist_test.txt\"\n",
        "TEST_BATCH_SIZE             = 4\n",
        "TEST_INPUT_SIZE             = 416\n",
        "TEST_DATA_AUG               = False\n",
        "TEST_DECTECTED_IMAGE_PATH   = \"\"\n",
        "TEST_SCORE_THRESHOLD        = 0.3\n",
        "TEST_IOU_THRESHOLD          = 0.45\n",
        "YOLO_STRIDES                = [8, 16, 32]\n",
        "YOLO_IOU_LOSS_THRESH        = 0.5\n",
        "YOLO_ANCHOR_PER_SCALE       = 3\n",
        "YOLO_MAX_BBOX_PER_SCALE     = 100"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MQxHNRCvNY0K",
        "colab_type": "text"
      },
      "source": [
        "### Layers"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_ozuTqG6LtP6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Relu6(Layer):\n",
        "    def __init__(self):\n",
        "        super(self.__class__,self).__init__()\n",
        "        self.relu6 = tf.nn.relu6\n",
        "        \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        return self.relu6(inputs)\n",
        "class BatchNorm(Layer):\n",
        "    def __init__(self, scale=True, center=True):\n",
        "        super(self.__class__,self).__init__()\n",
        "        #self.bn = tf.keras.layers.BatchNormalization(scale=scale, center=center, trainable=True)\n",
        "        self.bn = BatchNormalization(scale=scale, center=center, trainable=True)\n",
        "\n",
        "    #@tf.function\n",
        "    def call(self, inputs, training=True):\n",
        "        return self.bn(inputs, training=training)\n",
        "class Convolution2D(Layer):\n",
        "    def __init__(self, filters, kernel_size, is_activation = True, is_batch_norm = True, downsample=False):\n",
        "        super(self.__class__,self).__init__()\n",
        "        \n",
        "        self.downsample = downsample\n",
        "        if self.downsample:\n",
        "            padding = 'valid'\n",
        "            strides = 2\n",
        "        else:\n",
        "            strides = 1\n",
        "            padding = 'same'\n",
        "        \n",
        "        self.is_activation = is_activation\n",
        "        self.is_batch_norm = is_batch_norm\n",
        "        self.zero_pad =  ZeroPadding2D(((1, 0), (1, 0)))\n",
        "        self.conv = tf.keras.layers.Conv2D(filters = filters, kernel_size = kernel_size, strides = strides, padding=padding, \n",
        "                                           use_bias=not is_batch_norm, kernel_regularizer=l2(0.0005),\n",
        "                                           kernel_initializer=tf.random_normal_initializer(stddev=0.01),\n",
        "                                           bias_initializer=tf.constant_initializer(0.))\n",
        "        \n",
        "        self.bn = BatchNorm()\n",
        "        self.act = LeakyReLU(alpha=0.1)\n",
        "        \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        \n",
        "        if self.downsample:\n",
        "            x = self.zero_pad(inputs)\n",
        "#             print(\" PAD SHAOE \",x.shape )\n",
        "            x = self.conv(x)\n",
        "        else:\n",
        "            x = self.conv(inputs)\n",
        "            \n",
        "        if self.is_batch_norm:\n",
        "            x = self.bn(x)\n",
        "        if self.is_activation:\n",
        "            x = self.act(x)\n",
        "        \n",
        "        return x\n",
        "class AveragePooling(Layer):\n",
        "    def __init__(self, pool_size):\n",
        "        super(self.__class__,self).__init__()\n",
        "        self.avgpool = tf.keras.layers.AveragePooling2D(pool_size=pool_size, padding=\"SAME\")\n",
        "        \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        \n",
        "        x = self.avgpool(inputs)\n",
        "         \n",
        "        return x\n",
        "class DenseLayer(Layer):\n",
        "    def __init__(self, units):\n",
        "        super(self.__class__,self).__init__()\n",
        "\n",
        "        self.dense = tf.keras.layers.Dense(units=units,\n",
        "                                           kernel_initializer=tf.random_normal_initializer(stddev=0.01))\n",
        "        \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        \n",
        "        x = self.dense(inputs)\n",
        "        \n",
        "        return x\n",
        "class FlattenLayer(Layer):\n",
        "    def __init__(self):\n",
        "        super(self.__class__,self).__init__()\n",
        "        self.flatten = tf.keras.layers.Flatten()\n",
        "        \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        x = self.flatten(inputs)\n",
        "        return x\n",
        "class Residual_Block(Layer):\n",
        "    def __init__(self, filters):\n",
        "        super(self.__class__,self).__init__()\n",
        "        self.conv1  = Convolution2D(filters =   filters, kernel_size = (1, 1))\n",
        "        self.conv2  = Convolution2D(filters = 2*filters, kernel_size = (3, 3))\n",
        "        self.add = tf.keras.layers.Add()\n",
        "\n",
        "    def call(self, inputs):\n",
        "        short_cut = inputs\n",
        "        x = self.conv1(inputs)\n",
        "        x = self.conv2(x)\n",
        "        \n",
        "#         print(\" Residual Output 1\", short_cut.shape, x.shape)\n",
        "        residual_output = self.add([short_cut,x])\n",
        "#         print(\" Residual Output 2\", residual_output.shape)\n",
        "    \n",
        "        return residual_output\n",
        "class Upsample(Layer):\n",
        "    def __init__(self):\n",
        "        super(self.__class__,self).__init__()\n",
        "        \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        return tf.image.resize(inputs, (inputs.shape[1] * 2, inputs.shape[2] * 2), method='nearest')"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Rr0tuDcHLtbD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Darknet(Model):\n",
        "    def __init__(self):\n",
        "        super(self.__class__,self).__init__()\n",
        "        self.conv1        = Convolution2D(32, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv2        = Convolution2D(64, (3, 3), is_activation=True, is_batch_norm=True, downsample=True)\n",
        "        \n",
        "        self.residual1_1  = Residual_Block(32)\n",
        "        \n",
        "        self.conv3        = Convolution2D(128, (3, 3), is_activation=True, is_batch_norm=True, downsample=True)\n",
        "        \n",
        "        self.residual2_1  = Residual_Block(64)\n",
        "        self.residual2_2  = Residual_Block(64)\n",
        "        \n",
        "        self.conv4        = Convolution2D(256, (3, 3),is_activation=True, is_batch_norm=True, downsample=True)\n",
        "        \n",
        "        self.residual3_1  = Residual_Block(128)\n",
        "        self.residual3_2  = Residual_Block(128)\n",
        "        self.residual3_3  = Residual_Block(128)\n",
        "        self.residual3_4  = Residual_Block(128)\n",
        "        self.residual3_5  = Residual_Block(128)\n",
        "        self.residual3_6  = Residual_Block(128)\n",
        "        self.residual3_7  = Residual_Block(128)\n",
        "        self.residual3_8  = Residual_Block(128)\n",
        "        \n",
        "        self.conv5        = Convolution2D(512,  (3, 3),is_activation=True, is_batch_norm=True, downsample=True)\n",
        "        \n",
        "        self.residual4_1  = Residual_Block(256)\n",
        "        self.residual4_2  = Residual_Block(256)\n",
        "        self.residual4_3  = Residual_Block(256)\n",
        "        self.residual4_4  = Residual_Block(256)\n",
        "        self.residual4_5  = Residual_Block(256)\n",
        "        self.residual4_6  = Residual_Block(256)\n",
        "        self.residual4_7  = Residual_Block(256)\n",
        "        self.residual4_8  = Residual_Block(256)\n",
        "        \n",
        "        self.conv6        = Convolution2D(1024, (3, 3),is_activation=True, is_batch_norm=True, downsample=True)\n",
        "        \n",
        "        self.residual5_1  = Residual_Block(512)\n",
        "        self.residual5_2  = Residual_Block(512)\n",
        "        self.residual5_3  = Residual_Block(512)\n",
        "        self.residual5_4  = Residual_Block(512)\n",
        "             \n",
        "    @tf.function\n",
        "    def call(self, inputs):\n",
        "        x = self.conv1(inputs)\n",
        "        x = self.conv2(x) \n",
        "        \n",
        "        x = self.residual1_1(x)        \n",
        "        \n",
        "        x = self.conv3(x) \n",
        "        x = self.residual2_1(x)\n",
        "        x = self.residual2_2(x)\n",
        "        \n",
        "        x = self.conv4(x)\n",
        "        x = self.residual3_1(x)\n",
        "        x = self.residual3_2(x)\n",
        "        x = self.residual3_3(x)\n",
        "        x = self.residual3_4(x)\n",
        "        x = self.residual3_5(x)\n",
        "        x = self.residual3_6(x)\n",
        "        x = self.residual3_7(x)\n",
        "        x = self.residual3_8(x)\n",
        "        route1 = x\n",
        "        \n",
        "        x = self.conv5(x)\n",
        "        x = self.residual4_1(x)\n",
        "        x = self.residual4_2(x)\n",
        "        x = self.residual4_3(x)\n",
        "        x = self.residual4_4(x)\n",
        "        x = self.residual4_5(x)\n",
        "        x = self.residual4_6(x)\n",
        "        x = self.residual4_7(x)\n",
        "        x = self.residual4_8(x)\n",
        "        route2 = x\n",
        "        \n",
        "        x = self.conv6(x)\n",
        "        x = self.residual5_1(x)\n",
        "        x = self.residual5_2(x)\n",
        "        x = self.residual5_3(x)\n",
        "        x = self.residual5_4(x)\n",
        "        \n",
        "    \n",
        "        return route1, route2, x"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NxLPUSN-NdzB",
        "colab_type": "text"
      },
      "source": [
        "### Loss functions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ct_87nJ5Ltkv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def decode(conv_output, NUM_CLASS, i=0):\n",
        "    # where i = 0, 1 or 2 to correspond to the three grid scales  \n",
        "    conv_shape       = tf.shape(conv_output)\n",
        "    batch_size       = conv_shape[0]\n",
        "    output_size      = conv_shape[1]\n",
        "\n",
        "    conv_output = tf.reshape(conv_output, (batch_size, output_size, output_size, 3, 5 + NUM_CLASS))\n",
        "\n",
        "    conv_raw_dxdy = conv_output[:, :, :, :, 0:2] # offset of center position     \n",
        "    conv_raw_dwdh = conv_output[:, :, :, :, 2:4] # Prediction box length and width offset\n",
        "    conv_raw_conf = conv_output[:, :, :, :, 4:5] # confidence of the prediction box\n",
        "    conv_raw_prob = conv_output[:, :, :, :, 5: ] # category probability of the prediction box \n",
        "\n",
        "    # next need Draw the grid. Where output_size is equal to 13, 26 or 52  \n",
        "    y = tf.range(output_size, dtype=tf.int32)\n",
        "    y = tf.expand_dims(y, -1)\n",
        "    y = tf.tile(y, [1, output_size])\n",
        "    x = tf.range(output_size,dtype=tf.int32)\n",
        "    x = tf.expand_dims(x, 0)\n",
        "    x = tf.tile(x, [output_size, 1])\n",
        "\n",
        "    xy_grid = tf.concat([x[:, :, tf.newaxis], y[:, :, tf.newaxis]], axis=-1)\n",
        "    xy_grid = tf.tile(xy_grid[tf.newaxis, :, :, tf.newaxis, :], [batch_size, 1, 1, 3, 1])\n",
        "    xy_grid = tf.cast(xy_grid, tf.float32)\n",
        "\n",
        "    # Calculate the center position of the prediction box:\n",
        "    pred_xy = (tf.sigmoid(conv_raw_dxdy) + xy_grid) * STRIDES[i]\n",
        "    # Calculate the length and width of the prediction box:\n",
        "    pred_wh = (tf.exp(conv_raw_dwdh) * ANCHORS[i]) * STRIDES[i]\n",
        "\n",
        "    pred_xywh = tf.concat([pred_xy, pred_wh], axis=-1)\n",
        "    pred_conf = tf.sigmoid(conv_raw_conf) # object box calculates the predicted confidence\n",
        "    pred_prob = tf.sigmoid(conv_raw_prob) # calculating the predicted probability category box object\n",
        "\n",
        "    # calculating the predicted probability category box object\n",
        "    return tf.concat([pred_xywh, pred_conf, pred_prob], axis=-1)\n",
        "def bbox_iou(boxes1, boxes2):\n",
        "    boxes1_area = boxes1[..., 2] * boxes1[..., 3]\n",
        "    boxes2_area = boxes2[..., 2] * boxes2[..., 3]\n",
        "\n",
        "    boxes1 = tf.concat([boxes1[..., :2] - boxes1[..., 2:] * 0.5,\n",
        "                        boxes1[..., :2] + boxes1[..., 2:] * 0.5], axis=-1)\n",
        "    boxes2 = tf.concat([boxes2[..., :2] - boxes2[..., 2:] * 0.5,\n",
        "                        boxes2[..., :2] + boxes2[..., 2:] * 0.5], axis=-1)\n",
        "\n",
        "    left_up = tf.maximum(boxes1[..., :2], boxes2[..., :2])\n",
        "    right_down = tf.minimum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "\n",
        "    inter_section = tf.maximum(right_down - left_up, 0.0)\n",
        "    inter_area = inter_section[..., 0] * inter_section[..., 1]\n",
        "    union_area = boxes1_area + boxes2_area - inter_area\n",
        "\n",
        "    return 1.0 * inter_area / union_area\n",
        "def bbox_giou(boxes1, boxes2):\n",
        "    boxes1 = tf.concat([boxes1[..., :2] - boxes1[..., 2:] * 0.5,\n",
        "                        boxes1[..., :2] + boxes1[..., 2:] * 0.5], axis=-1)\n",
        "    boxes2 = tf.concat([boxes2[..., :2] - boxes2[..., 2:] * 0.5,\n",
        "                        boxes2[..., :2] + boxes2[..., 2:] * 0.5], axis=-1)\n",
        "\n",
        "    boxes1 = tf.concat([tf.minimum(boxes1[..., :2], boxes1[..., 2:]),\n",
        "                        tf.maximum(boxes1[..., :2], boxes1[..., 2:])], axis=-1)\n",
        "    boxes2 = tf.concat([tf.minimum(boxes2[..., :2], boxes2[..., 2:]),\n",
        "                        tf.maximum(boxes2[..., :2], boxes2[..., 2:])], axis=-1)\n",
        "\n",
        "    boxes1_area = (boxes1[..., 2] - boxes1[..., 0]) * (boxes1[..., 3] - boxes1[..., 1])\n",
        "    boxes2_area = (boxes2[..., 2] - boxes2[..., 0]) * (boxes2[..., 3] - boxes2[..., 1])\n",
        "\n",
        "    left_up = tf.maximum(boxes1[..., :2], boxes2[..., :2])\n",
        "    right_down = tf.minimum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "\n",
        "    inter_section = tf.maximum(right_down - left_up, 0.0)\n",
        "    inter_area = inter_section[..., 0] * inter_section[..., 1]\n",
        "    union_area = boxes1_area + boxes2_area - inter_area\n",
        "\n",
        "    # Calculate the iou value between the two bounding boxes\n",
        "    iou = inter_area / union_area\n",
        "\n",
        "    # Calculate the coordinates of the upper left corner and the lower right corner of the smallest closed convex surface\n",
        "    enclose_left_up = tf.minimum(boxes1[..., :2], boxes2[..., :2])\n",
        "    enclose_right_down = tf.maximum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "    enclose = tf.maximum(enclose_right_down - enclose_left_up, 0.0)\n",
        "\n",
        "    # Calculate the area of the smallest closed convex surface C\n",
        "    enclose_area = enclose[..., 0] * enclose[..., 1]\n",
        "\n",
        "    # Calculate the GIoU value according to the GioU formula  \n",
        "    giou = iou - 1.0 * (enclose_area - union_area) / enclose_area\n",
        "\n",
        "    return giou\n",
        "# testing (should be better than giou)\n",
        "def bbox_ciou(boxes1, boxes2):\n",
        "    boxes1_coor = tf.concat([boxes1[..., :2] - boxes1[..., 2:] * 0.5,\n",
        "                        boxes1[..., :2] + boxes1[..., 2:] * 0.5], axis=-1)\n",
        "    boxes2_coor = tf.concat([boxes2[..., :2] - boxes2[..., 2:] * 0.5,\n",
        "                        boxes2[..., :2] + boxes2[..., 2:] * 0.5], axis=-1)\n",
        "\n",
        "    left = tf.maximum(boxes1_coor[..., 0], boxes2_coor[..., 0])\n",
        "    up = tf.maximum(boxes1_coor[..., 1], boxes2_coor[..., 1])\n",
        "    right = tf.maximum(boxes1_coor[..., 2], boxes2_coor[..., 2])\n",
        "    down = tf.maximum(boxes1_coor[..., 3], boxes2_coor[..., 3])\n",
        "\n",
        "    c = (right - left) * (right - left) + (up - down) * (up - down)\n",
        "    iou = bbox_iou(boxes1, boxes2)\n",
        "\n",
        "    u = (boxes1[..., 0] - boxes2[..., 0]) * (boxes1[..., 0] - boxes2[..., 0]) + (boxes1[..., 1] - boxes2[..., 1]) * (boxes1[..., 1] - boxes2[..., 1])\n",
        "    d = u / c\n",
        "\n",
        "    ar_gt = boxes2[..., 2] / boxes2[..., 3]\n",
        "    ar_pred = boxes1[..., 2] / boxes1[..., 3]\n",
        "\n",
        "    ar_loss = 4 / (np.pi * np.pi) * (tf.atan(ar_gt) - tf.atan(ar_pred)) * (tf.atan(ar_gt) - tf.atan(ar_pred))\n",
        "    alpha = ar_loss / (1 - iou + ar_loss + 0.000001)\n",
        "    ciou_term = d + alpha * ar_loss\n",
        "\n",
        "    return iou - ciou_term\n",
        "def compute_loss(pred, conv, label, bboxes, i=0, CLASSES=TRAIN_CLASSES):\n",
        "    NUM_CLASS = len(read_class_names(CLASSES))\n",
        "    conv_shape  = tf.shape(conv)\n",
        "    batch_size  = conv_shape[0]\n",
        "    output_size = conv_shape[1]\n",
        "    input_size  = STRIDES[i] * output_size\n",
        "    conv = tf.reshape(conv, (batch_size, output_size, output_size, 3, 5 + NUM_CLASS))\n",
        "\n",
        "    conv_raw_conf = conv[:, :, :, :, 4:5]\n",
        "    conv_raw_prob = conv[:, :, :, :, 5:]\n",
        "\n",
        "    pred_xywh     = pred[:, :, :, :, 0:4]\n",
        "    pred_conf     = pred[:, :, :, :, 4:5]\n",
        "\n",
        "    label_xywh    = label[:, :, :, :, 0:4]\n",
        "    respond_bbox  = label[:, :, :, :, 4:5]\n",
        "    label_prob    = label[:, :, :, :, 5:]\n",
        "\n",
        "    giou = tf.expand_dims(bbox_giou(pred_xywh, label_xywh), axis=-1)\n",
        "    input_size = tf.cast(input_size, tf.float32)\n",
        "\n",
        "    bbox_loss_scale = 2.0 - 1.0 * label_xywh[:, :, :, :, 2:3] * label_xywh[:, :, :, :, 3:4] / (input_size ** 2)\n",
        "    giou_loss = respond_bbox * bbox_loss_scale * (1 - giou)\n",
        "\n",
        "    iou = bbox_iou(pred_xywh[:, :, :, :, np.newaxis, :], bboxes[:, np.newaxis, np.newaxis, np.newaxis, :, :])\n",
        "    # Find the value of IoU with the real box The largest prediction box\n",
        "    max_iou = tf.expand_dims(tf.reduce_max(iou, axis=-1), axis=-1)\n",
        "\n",
        "    # If the largest iou is less than the threshold, it is considered that the prediction box contains no objects, then the background box\n",
        "    respond_bgd = (1.0 - respond_bbox) * tf.cast( max_iou < YOLO_IOU_LOSS_THRESH, tf.float32 )\n",
        "\n",
        "    conf_focal = tf.pow(respond_bbox - pred_conf, 2)\n",
        "\n",
        "    # Calculate the loss of confidence\n",
        "    # we hope that if the grid contains objects, then the network output prediction box has a confidence of 1 and 0 when there is no object.\n",
        "    conf_loss = conf_focal * (\n",
        "            respond_bbox * tf.nn.sigmoid_cross_entropy_with_logits(labels=respond_bbox, logits=conv_raw_conf)\n",
        "            +\n",
        "            respond_bgd * tf.nn.sigmoid_cross_entropy_with_logits(labels=respond_bbox, logits=conv_raw_conf)\n",
        "    )\n",
        "\n",
        "    prob_loss = respond_bbox * tf.nn.sigmoid_cross_entropy_with_logits(labels=label_prob, logits=conv_raw_prob)\n",
        "\n",
        "    giou_loss = tf.reduce_mean(tf.reduce_sum(giou_loss, axis=[1,2,3,4]))\n",
        "    conf_loss = tf.reduce_mean(tf.reduce_sum(conf_loss, axis=[1,2,3,4]))\n",
        "    prob_loss = tf.reduce_mean(tf.reduce_sum(prob_loss, axis=[1,2,3,4]))\n",
        "\n",
        "    return giou_loss, conf_loss, prob_loss"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DcwyCu9uLtpI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def read_class_names(class_file_name):\n",
        "    # loads class name from a file\n",
        "    names = {}\n",
        "    with open(class_file_name, 'r') as data:\n",
        "        for ID, name in enumerate(data):\n",
        "            names[ID] = name.strip('\\n')\n",
        "    return names"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uP_FzHJuNjW7",
        "colab_type": "text"
      },
      "source": [
        "### Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UrmaROr1LtgI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class YOLOV3(Model):\n",
        "    def __init__(self, NUM_CLASS, TRAIN = False):\n",
        "        super(self.__class__,self).__init__()\n",
        "        self.NUM_CLASS = NUM_CLASS\n",
        "        self.TRAIN = TRAIN\n",
        "        self.darknet = Darknet()\n",
        "        self.conv1 = Convolution2D(512,  (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv2 = Convolution2D(1024, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv3 = Convolution2D(512,  (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv4 = Convolution2D(1024, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv5 = Convolution2D(512,  (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        \n",
        "        self.conv_lobj_branch = Convolution2D(1024, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv_lbbox = Convolution2D(3*(NUM_CLASS + 5),  (1, 1), is_activation=False, is_batch_norm=False, downsample=False)\n",
        "        \n",
        "        self.conv6 = Convolution2D(256,  (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.upsample1 = Upsample()\n",
        "         \n",
        "        self.conv7  = Convolution2D(256, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv8  = Convolution2D(512, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv9  = Convolution2D(256, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv10 = Convolution2D(512, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv11 = Convolution2D(256, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        \n",
        "        \n",
        "        self.conv_mobj_branch = Convolution2D(512, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv_mbbox = Convolution2D(3*(NUM_CLASS + 5), (1, 1), is_activation=False, is_batch_norm=False, downsample=False)\n",
        "        #convolutional(conv_mobj_branch, (1, 1, 512, 3*(NUM_CLASS + 5)), activate=False, bn=False)\n",
        "\n",
        "        \n",
        "        self.conv12 = Convolution2D(128, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.upsample2  = Upsample()\n",
        "        self.conv13 = Convolution2D(128, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv14 = Convolution2D(256, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv15 = Convolution2D(128, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv16 = Convolution2D(256, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "        self.conv17 = Convolution2D(128, (1, 1), is_activation=True, is_batch_norm=True, downsample=False)    \n",
        "\n",
        "        self.conv_sobj_branch = Convolution2D(256, (3, 3), is_activation=True, is_batch_norm=True, downsample=False)\n",
        "    # conv_sbbox is used to predict small size objects, shape = [None, 52, 52, 255]\n",
        "        self.conv_sbbox = Convolution2D(3*(NUM_CLASS +5), (1, 1), is_activation=False, is_batch_norm=False, downsample=False)\n",
        "\n",
        "    @tf.function\n",
        "    def call(self, inputs, training=True):\n",
        "        \n",
        "        route_1, route_2, conv  = self.darknet(inputs)    \n",
        "        x = self.conv1(conv)\n",
        "        x = self.conv2(x)\n",
        "        x = self.conv3(x)\n",
        "        x = self.conv4(x)\n",
        "        x = self.conv5(x)\n",
        "        conv_lobj_branch = self.conv_lobj_branch(x)\n",
        "        conv_lbbox = self.conv_lbbox(conv_lobj_branch)\n",
        "        \n",
        "        x = self.conv6(x)\n",
        "        x = self.upsample1(x)\n",
        "        \n",
        "        x = tf.concat([x, route_2], axis=-1)\n",
        "        x = self.conv7(x)\n",
        "        x = self.conv8(x)\n",
        "        x = self.conv9(x)\n",
        "        x = self.conv10(x)\n",
        "        x = self.conv11(x)\n",
        "        \n",
        "        conv_mobj_branch = self.conv_mobj_branch(x)\n",
        "        conv_mbbox = self.conv_mbbox(conv_mobj_branch)\n",
        "        \n",
        "        x = self.conv12(x)\n",
        "        x = self.upsample2(x)\n",
        "        x = tf.concat([x, route_1], axis=-1)\n",
        "        x = self.conv13(x)\n",
        "        x = self.conv14(x)\n",
        "        x = self.conv15(x)\n",
        "        x = self.conv16(x)\n",
        "        x = self.conv17(x)\n",
        "        conv_sobj_branch = self.conv_sobj_branch(x)\n",
        "        conv_sbbox = self.conv_sbbox(conv_sobj_branch)\n",
        "        \n",
        "        conv_tensors = [conv_sbbox, conv_mbbox, conv_lbbox]\n",
        "        \n",
        "        output_tensors = []\n",
        "\n",
        "        for i, conv_tensor in enumerate(conv_tensors):\n",
        "            pred_tensor = decode(conv_tensor, self.NUM_CLASS, i)\n",
        "            if self.TRAIN:\n",
        "                output_tensors.append(conv_tensor)\n",
        "            output_tensors.append(pred_tensor)\n",
        "        \n",
        "        return output_tensors"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rhBD26H-Ltdm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "yolo = YOLOV3(10, TRAIN=True)\n",
        "tert = np.random.rand(1,416,416,3).astype(np.float32)\n",
        "bert = yolo(tert)"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uOqcfL0tLtVC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "8f9ee7cc-d840-4685-ccbf-ec92cf8c7c5c"
      },
      "source": [
        "bert[0].shape, bert[1].shape, bert[2].shape"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(TensorShape([1, 52, 52, 45]),\n",
              " TensorShape([1, 52, 52, 3, 15]),\n",
              " TensorShape([1, 26, 26, 45]))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YFRx45CjLtUA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w2QS7yUzNmUm",
        "colab_type": "text"
      },
      "source": [
        "### Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2Hii1hAoLtBi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7iSXwJihLtGr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from yolov3.dataset import Dataset\n",
        "trainset = Dataset('train')\n",
        "testset = Dataset('test')"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_0cmnazCMA_4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NobVbkQ-MBTU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "steps_per_epoch = len(trainset)\n",
        "global_steps = tf.Variable(1, trainable=False, dtype=tf.int64)\n",
        "warmup_steps = TRAIN_WARMUP_EPOCHS * steps_per_epoch\n",
        "total_steps = TRAIN_EPOCHS * steps_per_epoch"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rjYwDz2MpkTy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ybcjjd5PpknP",
        "colab_type": "text"
      },
      "source": [
        "### debug"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EtX0pRVCpQJy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "YOLO_COCO_CLASSES = \"model_data/coco/coco.names\"\n",
        "def decode(conv_output, NUM_CLASS, i=0):\n",
        "    # where i = 0, 1 or 2 to correspond to the three grid scales  \n",
        "    conv_shape       = tf.shape(conv_output)\n",
        "    batch_size       = conv_shape[0]\n",
        "    output_size      = conv_shape[1]\n",
        "\n",
        "    conv_output = tf.reshape(conv_output, (batch_size, output_size, output_size, 3, 5 + NUM_CLASS))\n",
        "\n",
        "    conv_raw_dxdy = conv_output[:, :, :, :, 0:2] # offset of center position     \n",
        "    conv_raw_dwdh = conv_output[:, :, :, :, 2:4] # Prediction box length and width offset\n",
        "    conv_raw_conf = conv_output[:, :, :, :, 4:5] # confidence of the prediction box\n",
        "    conv_raw_prob = conv_output[:, :, :, :, 5: ] # category probability of the prediction box \n",
        "\n",
        "    # next need Draw the grid. Where output_size is equal to 13, 26 or 52  \n",
        "    y = tf.range(output_size, dtype=tf.int32)\n",
        "    y = tf.expand_dims(y, -1)\n",
        "    y = tf.tile(y, [1, output_size])\n",
        "    x = tf.range(output_size,dtype=tf.int32)\n",
        "    x = tf.expand_dims(x, 0)\n",
        "    x = tf.tile(x, [output_size, 1])\n",
        "\n",
        "    xy_grid = tf.concat([x[:, :, tf.newaxis], y[:, :, tf.newaxis]], axis=-1)\n",
        "    xy_grid = tf.tile(xy_grid[tf.newaxis, :, :, tf.newaxis, :], [batch_size, 1, 1, 3, 1])\n",
        "    xy_grid = tf.cast(xy_grid, tf.float32)\n",
        "\n",
        "    # Calculate the center position of the prediction box:\n",
        "    pred_xy = (tf.sigmoid(conv_raw_dxdy) + xy_grid) * STRIDES[i]\n",
        "    # Calculate the length and width of the prediction box:\n",
        "    pred_wh = (tf.exp(conv_raw_dwdh) * ANCHORS[i]) * STRIDES[i]\n",
        "\n",
        "    pred_xywh = tf.concat([pred_xy, pred_wh], axis=-1)\n",
        "    pred_conf = tf.sigmoid(conv_raw_conf) # object box calculates the predicted confidence\n",
        "    pred_prob = tf.sigmoid(conv_raw_prob) # calculating the predicted probability category box object\n",
        "\n",
        "    # calculating the predicted probability category box object\n",
        "    return tf.concat([pred_xywh, pred_conf, pred_prob], axis=-1)\n",
        "\n",
        "def bbox_iou(boxes1, boxes2):\n",
        "    boxes1_area = boxes1[..., 2] * boxes1[..., 3]\n",
        "    boxes2_area = boxes2[..., 2] * boxes2[..., 3]\n",
        "\n",
        "    boxes1 = tf.concat([boxes1[..., :2] - boxes1[..., 2:] * 0.5,\n",
        "                        boxes1[..., :2] + boxes1[..., 2:] * 0.5], axis=-1)\n",
        "    boxes2 = tf.concat([boxes2[..., :2] - boxes2[..., 2:] * 0.5,\n",
        "                        boxes2[..., :2] + boxes2[..., 2:] * 0.5], axis=-1)\n",
        "\n",
        "    left_up = tf.maximum(boxes1[..., :2], boxes2[..., :2])\n",
        "    right_down = tf.minimum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "\n",
        "    inter_section = tf.maximum(right_down - left_up, 0.0)\n",
        "    inter_area = inter_section[..., 0] * inter_section[..., 1]\n",
        "    union_area = boxes1_area + boxes2_area - inter_area\n",
        "\n",
        "    return 1.0 * inter_area / union_area\n",
        "\n",
        "def bbox_giou(boxes1, boxes2):\n",
        "    boxes1 = tf.concat([boxes1[..., :2] - boxes1[..., 2:] * 0.5,\n",
        "                        boxes1[..., :2] + boxes1[..., 2:] * 0.5], axis=-1)\n",
        "    boxes2 = tf.concat([boxes2[..., :2] - boxes2[..., 2:] * 0.5,\n",
        "                        boxes2[..., :2] + boxes2[..., 2:] * 0.5], axis=-1)\n",
        "\n",
        "    boxes1 = tf.concat([tf.minimum(boxes1[..., :2], boxes1[..., 2:]),\n",
        "                        tf.maximum(boxes1[..., :2], boxes1[..., 2:])], axis=-1)\n",
        "    boxes2 = tf.concat([tf.minimum(boxes2[..., :2], boxes2[..., 2:]),\n",
        "                        tf.maximum(boxes2[..., :2], boxes2[..., 2:])], axis=-1)\n",
        "\n",
        "    boxes1_area = (boxes1[..., 2] - boxes1[..., 0]) * (boxes1[..., 3] - boxes1[..., 1])\n",
        "    boxes2_area = (boxes2[..., 2] - boxes2[..., 0]) * (boxes2[..., 3] - boxes2[..., 1])\n",
        "\n",
        "    left_up = tf.maximum(boxes1[..., :2], boxes2[..., :2])\n",
        "    right_down = tf.minimum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "\n",
        "    inter_section = tf.maximum(right_down - left_up, 0.0)\n",
        "    inter_area = inter_section[..., 0] * inter_section[..., 1]\n",
        "    union_area = boxes1_area + boxes2_area - inter_area\n",
        "\n",
        "    # Calculate the iou value between the two bounding boxes\n",
        "    iou = inter_area / union_area\n",
        "\n",
        "    # Calculate the coordinates of the upper left corner and the lower right corner of the smallest closed convex surface\n",
        "    enclose_left_up = tf.minimum(boxes1[..., :2], boxes2[..., :2])\n",
        "    enclose_right_down = tf.maximum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "    enclose = tf.maximum(enclose_right_down - enclose_left_up, 0.0)\n",
        "\n",
        "    # Calculate the area of the smallest closed convex surface C\n",
        "    enclose_area = enclose[..., 0] * enclose[..., 1]\n",
        "\n",
        "    # Calculate the GIoU value according to the GioU formula  \n",
        "    giou = iou - 1.0 * (enclose_area - union_area) / enclose_area\n",
        "\n",
        "    return giou\n",
        "\n",
        "# testing (should be better than giou)\n",
        "def bbox_ciou(boxes1, boxes2):\n",
        "    boxes1_coor = tf.concat([boxes1[..., :2] - boxes1[..., 2:] * 0.5,\n",
        "                        boxes1[..., :2] + boxes1[..., 2:] * 0.5], axis=-1)\n",
        "    boxes2_coor = tf.concat([boxes2[..., :2] - boxes2[..., 2:] * 0.5,\n",
        "                        boxes2[..., :2] + boxes2[..., 2:] * 0.5], axis=-1)\n",
        "\n",
        "    left = tf.maximum(boxes1_coor[..., 0], boxes2_coor[..., 0])\n",
        "    up = tf.maximum(boxes1_coor[..., 1], boxes2_coor[..., 1])\n",
        "    right = tf.maximum(boxes1_coor[..., 2], boxes2_coor[..., 2])\n",
        "    down = tf.maximum(boxes1_coor[..., 3], boxes2_coor[..., 3])\n",
        "\n",
        "    c = (right - left) * (right - left) + (up - down) * (up - down)\n",
        "    iou = bbox_iou(boxes1, boxes2)\n",
        "\n",
        "    u = (boxes1[..., 0] - boxes2[..., 0]) * (boxes1[..., 0] - boxes2[..., 0]) + (boxes1[..., 1] - boxes2[..., 1]) * (boxes1[..., 1] - boxes2[..., 1])\n",
        "    d = u / c\n",
        "\n",
        "    ar_gt = boxes2[..., 2] / boxes2[..., 3]\n",
        "    ar_pred = boxes1[..., 2] / boxes1[..., 3]\n",
        "\n",
        "    ar_loss = 4 / (np.pi * np.pi) * (tf.atan(ar_gt) - tf.atan(ar_pred)) * (tf.atan(ar_gt) - tf.atan(ar_pred))\n",
        "    alpha = ar_loss / (1 - iou + ar_loss + 0.000001)\n",
        "    ciou_term = d + alpha * ar_loss\n",
        "\n",
        "    return iou - ciou_term\n",
        "\n",
        "\n",
        "def compute_loss(pred, conv, label, bboxes, i=0, CLASSES=YOLO_COCO_CLASSES):\n",
        "    NUM_CLASS = len(read_class_names(CLASSES))\n",
        "    conv_shape  = tf.shape(conv)\n",
        "    batch_size  = conv_shape[0]\n",
        "    output_size = conv_shape[1]\n",
        "    input_size  = STRIDES[i] * output_size\n",
        "    conv = tf.reshape(conv, (batch_size, output_size, output_size, 3, 5 + NUM_CLASS))\n",
        "\n",
        "    conv_raw_conf = conv[:, :, :, :, 4:5]\n",
        "    conv_raw_prob = conv[:, :, :, :, 5:]\n",
        "\n",
        "    pred_xywh     = pred[:, :, :, :, 0:4]\n",
        "    pred_conf     = pred[:, :, :, :, 4:5]\n",
        "\n",
        "    label_xywh    = label[:, :, :, :, 0:4]\n",
        "    respond_bbox  = label[:, :, :, :, 4:5]\n",
        "    label_prob    = label[:, :, :, :, 5:]\n",
        "\n",
        "    giou = tf.expand_dims(bbox_giou(pred_xywh, label_xywh), axis=-1)\n",
        "    input_size = tf.cast(input_size, tf.float32)\n",
        "\n",
        "    bbox_loss_scale = 2.0 - 1.0 * label_xywh[:, :, :, :, 2:3] * label_xywh[:, :, :, :, 3:4] / (input_size ** 2)\n",
        "    giou_loss = respond_bbox * bbox_loss_scale * (1 - giou)\n",
        "\n",
        "    iou = bbox_iou(pred_xywh[:, :, :, :, np.newaxis, :], bboxes[:, np.newaxis, np.newaxis, np.newaxis, :, :])\n",
        "    # Find the value of IoU with the real box The largest prediction box\n",
        "    max_iou = tf.expand_dims(tf.reduce_max(iou, axis=-1), axis=-1)\n",
        "\n",
        "    # If the largest iou is less than the threshold, it is considered that the prediction box contains no objects, then the background box\n",
        "    respond_bgd = (1.0 - respond_bbox) * tf.cast( max_iou < YOLO_IOU_LOSS_THRESH, tf.float32 )\n",
        "\n",
        "    conf_focal = tf.pow(respond_bbox - pred_conf, 2)\n",
        "\n",
        "    # Calculate the loss of confidence\n",
        "    # we hope that if the grid contains objects, then the network output prediction box has a confidence of 1 and 0 when there is no object.\n",
        "    conf_loss = conf_focal * (\n",
        "            respond_bbox * tf.nn.sigmoid_cross_entropy_with_logits(labels=respond_bbox, logits=conv_raw_conf)\n",
        "            +\n",
        "            respond_bgd * tf.nn.sigmoid_cross_entropy_with_logits(labels=respond_bbox, logits=conv_raw_conf)\n",
        "    )\n",
        "\n",
        "    prob_loss = respond_bbox * tf.nn.sigmoid_cross_entropy_with_logits(labels=label_prob, logits=conv_raw_prob)\n",
        "\n",
        "    giou_loss = tf.reduce_mean(tf.reduce_sum(giou_loss, axis=[1,2,3,4]))\n",
        "    conf_loss = tf.reduce_mean(tf.reduce_sum(conf_loss, axis=[1,2,3,4]))\n",
        "    prob_loss = tf.reduce_mean(tf.reduce_sum(prob_loss, axis=[1,2,3,4]))\n",
        "\n",
        "    return giou_loss, conf_loss, prob_loss\n"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MsE2vBb9NoNL",
        "colab_type": "text"
      },
      "source": [
        "### Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WOsdtVmnJnxl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp /content/drive/My\\ Drive/delete/YOLO/CHECK4.zip ."
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hRpq5nkQJ9Ao",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        },
        "outputId": "6b9dde3c-97e6-41d6-93a7-c03988bdda5c"
      },
      "source": [
        "!unzip CHECK4.zip"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Archive:  CHECK4.zip\n",
            "   creating: CHECK4/\n",
            "  inflating: CHECK4/checkpoint       \n",
            "  inflating: CHECK4/check4.data-00000-of-00001  \n",
            "  inflating: CHECK4/check4.index     \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R8MQC1e9vVcT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "a1a5be54-1631-4097-bd7b-9bd59381fe3e"
      },
      "source": [
        "yolo.load_weights(\"CHECK4/check4\")"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7f8b31442048>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BTBUxWTMMBhp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "if os.path.exists(TRAIN_LOGDIR): shutil.rmtree(TRAIN_LOGDIR)\n",
        "writer = tf.summary.create_file_writer(TRAIN_LOGDIR)\n",
        "\n",
        "validate_writer = tf.summary.create_file_writer(TRAIN_LOGDIR)\n",
        "def validate_step(image_data, target):\n",
        "    with tf.GradientTape() as tape:\n",
        "        pred_result = yolo(image_data, training=False)\n",
        "        giou_loss=conf_loss=prob_loss=0\n",
        "\n",
        "        # optimizing process\n",
        "        grid = 3 if not TRAIN_YOLO_TINY else 2\n",
        "        for i in range(grid):\n",
        "            conv, pred = pred_result[i*2], pred_result[i*2+1]\n",
        "            loss_items = compute_loss(pred, conv, *target[i], i, CLASSES=TRAIN_CLASSES)\n",
        "            giou_loss += loss_items[0]\n",
        "            conf_loss += loss_items[1]\n",
        "            prob_loss += loss_items[2]\n",
        "\n",
        "        total_loss = giou_loss + conf_loss + prob_loss\n",
        "\n",
        "    return giou_loss.numpy(), conf_loss.numpy(), prob_loss.numpy(), total_loss.numpy()"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n_d6MqFzMBc8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "optimizer = tf.keras.optimizers.Adam()\n",
        "\n",
        "def train_step(image_data, target):\n",
        "    with tf.GradientTape() as tape:\n",
        "        pred_result = yolo(image_data, training=True)\n",
        "        giou_loss=conf_loss=prob_loss=0\n",
        "\n",
        "        # optimizing process\n",
        "        grid = 3 if not TRAIN_YOLO_TINY else 2\n",
        "        for i in range(grid):\n",
        "            conv, pred = pred_result[i*2], pred_result[i*2+1]\n",
        "            loss_items = compute_loss(pred, conv, *target[i], i, CLASSES=TRAIN_CLASSES)\n",
        "            giou_loss += loss_items[0]\n",
        "            conf_loss += loss_items[1]\n",
        "            prob_loss += loss_items[2]\n",
        "\n",
        "        total_loss = giou_loss + conf_loss + prob_loss\n",
        "        print(\" Total Loss \", total_loss)\n",
        "\n",
        "        gradients = tape.gradient(total_loss, yolo.trainable_variables)\n",
        "        optimizer.apply_gradients(zip(gradients, yolo.trainable_variables))\n",
        "\n",
        "        # update learning rate\n",
        "        # about warmup: https://arxiv.org/pdf/1812.01187.pdf&usg=ALkJrhglKOPDjNt6SHGbphTHyMcT0cuMJg\n",
        "        global_steps.assign_add(1)\n",
        "        if global_steps < warmup_steps:# and not TRAIN_TRANSFER:\n",
        "            lr = global_steps / warmup_steps * TRAIN_LR_INIT\n",
        "        else:\n",
        "            lr = TRAIN_LR_END + 0.5 * (TRAIN_LR_INIT - TRAIN_LR_END)*(\n",
        "                (1 + tf.cos((global_steps - warmup_steps) / (total_steps - warmup_steps) * np.pi)))\n",
        "        optimizer.lr.assign(lr.numpy())\n",
        "\n",
        "        # writing summary data\n",
        "        with writer.as_default():\n",
        "            tf.summary.scalar(\"lr\", optimizer.lr, step=global_steps)\n",
        "            tf.summary.scalar(\"loss/total_loss\", total_loss, step=global_steps)\n",
        "            tf.summary.scalar(\"loss/giou_loss\", giou_loss, step=global_steps)\n",
        "            tf.summary.scalar(\"loss/conf_loss\", conf_loss, step=global_steps)\n",
        "            tf.summary.scalar(\"loss/prob_loss\", prob_loss, step=global_steps)\n",
        "        writer.flush()\n",
        "\n",
        "    return global_steps.numpy(), optimizer.lr.numpy(), giou_loss.numpy(), conf_loss.numpy(), prob_loss.numpy(), total_loss.numpy()"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KWokub8ue9BO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "TRAIN_SAVE_BEST_ONLY        = True # saves only best model according validation loss (True recommended)\n",
        "TRAIN_SAVE_CHECKPOINT       = False"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-CRg64o-gMAb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "save_directory = \"CHECK5/check5\""
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ull_tYXJm2BM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0908f620-8415-413f-cc43-b7050c383f87"
      },
      "source": [
        "%cd /content/TensorFlow-2.x-YOLOv3/"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/TensorFlow-2.x-YOLOv3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aOpK6r2QMBbm",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "4254721e-9f46-4113-848e-1af20df80dd4"
      },
      "source": [
        "best_val_loss = 1000 # should be large at start\n",
        "for epoch in range(TRAIN_EPOCHS):\n",
        "    for image_data, target in trainset:\n",
        "        results = train_step(image_data, target)\n",
        "        print(\" ****  \")\n",
        "        cur_step = results[0]%steps_per_epoch\n",
        "        print(\"epoch:{:2.0f} step:{:5.0f}/{}, lr:{:.6f}, giou_loss:{:7.2f}, conf_loss:{:7.2f}, prob_loss:{:7.2f}, total_loss:{:7.2f}\"\n",
        "              .format(epoch, cur_step, steps_per_epoch, results[1], results[2], results[3], results[4], results[5]))\n",
        "\n",
        "    \n",
        "    if len(testset) == 0:\n",
        "        print(\"configure TEST options to validate model\")\n",
        "        yolo.save_weights(os.path.join(TRAIN_CHECKPOINTS_FOLDER, TRAIN_MODEL_NAME))\n",
        "        continue\n",
        "\n",
        "    count, giou_val, conf_val, prob_val, total_val = 0., 0, 0, 0, 0\n",
        "    for image_data, target in testset:\n",
        "        results = validate_step(image_data, target)\n",
        "        count += 1\n",
        "        giou_val += results[0]\n",
        "        conf_val += results[1]\n",
        "        prob_val += results[2]\n",
        "        total_val += results[3]\n",
        "    # writing validate summary data\n",
        "    with validate_writer.as_default():\n",
        "        tf.summary.scalar(\"validate_loss/total_val\", total_val/count, step=epoch)\n",
        "        tf.summary.scalar(\"validate_loss/giou_val\", giou_val/count, step=epoch)\n",
        "        tf.summary.scalar(\"validate_loss/conf_val\", conf_val/count, step=epoch)\n",
        "        tf.summary.scalar(\"validate_loss/prob_val\", prob_val/count, step=epoch)\n",
        "    validate_writer.flush()\n",
        "\n",
        "    print(\"\\n\\ngiou_val_loss:{:7.2f}, conf_val_loss:{:7.2f}, prob_val_loss:{:7.2f}, total_val_loss:{:7.2f}\\n\\n\".\n",
        "          format(giou_val/count, conf_val/count, prob_val/count, total_val/count))\n",
        "\n",
        "    # if TRAIN_SAVE_CHECKPOINT and not TRAIN_SAVE_BEST_ONLY:\n",
        "    #     save_directory = os.path.join(TRAIN_CHECKPOINTS_FOLDER, TRAIN_MODEL_NAME+\"_val_loss_{:7.2f}\".format(total_val/count))\n",
        "    yolo.save_weights(save_directory)\n",
        "    # if TRAIN_SAVE_BEST_ONLY and best_val_loss>total_val/count:\n",
        "    #     save_directory = os.path.join(TRAIN_CHECKPOINTS_FOLDER, TRAIN_MODEL_NAME)\n",
        "    #     yolo.save_weights(save_directory)\n",
        "    #     best_val_loss = total_val/count\n",
        "    # if not TRAIN_SAVE_BEST_ONLY and not TRAIN_SAVE_CHECKPOINT:\n",
        "    #     save_directory = os.path.join(TRAIN_CHECKPOINTS_FOLDER, TRAIN_MODEL_NAME)\n",
        "    #     yolo.save_weights(save_directory)\n"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[1;30;43mStreaming output truncated to the last 5000 lines.\u001b[0m\n",
            " Total Loss  tf.Tensor(4.8361855, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  109/125, lr:0.000005, giou_loss:   2.32, conf_loss:   0.15, prob_loss:   2.36, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.7151513, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  110/125, lr:0.000005, giou_loss:   2.24, conf_loss:   0.06, prob_loss:   2.42, total_loss:   4.72\n",
            " Total Loss  tf.Tensor(4.422655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  111/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.19, prob_loss:   2.13, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(4.6632023, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  112/125, lr:0.000005, giou_loss:   2.18, conf_loss:   0.09, prob_loss:   2.39, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.1823754, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  113/125, lr:0.000005, giou_loss:   1.92, conf_loss:   0.10, prob_loss:   2.16, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(5.085829, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  114/125, lr:0.000005, giou_loss:   2.33, conf_loss:   0.16, prob_loss:   2.60, total_loss:   5.09\n",
            " Total Loss  tf.Tensor(4.7541165, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  115/125, lr:0.000005, giou_loss:   2.32, conf_loss:   0.10, prob_loss:   2.33, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(4.279336, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  116/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.08, prob_loss:   2.08, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(3.5219274, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  117/125, lr:0.000005, giou_loss:   1.63, conf_loss:   0.14, prob_loss:   1.75, total_loss:   3.52\n",
            " Total Loss  tf.Tensor(4.2996845, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  118/125, lr:0.000005, giou_loss:   1.97, conf_loss:   0.02, prob_loss:   2.31, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(3.51092, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  119/125, lr:0.000005, giou_loss:   1.63, conf_loss:   0.09, prob_loss:   1.79, total_loss:   3.51\n",
            " Total Loss  tf.Tensor(5.3579283, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  120/125, lr:0.000005, giou_loss:   2.39, conf_loss:   0.24, prob_loss:   2.72, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(5.656951, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  121/125, lr:0.000005, giou_loss:   2.69, conf_loss:   0.23, prob_loss:   2.74, total_loss:   5.66\n",
            " Total Loss  tf.Tensor(3.9040852, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  122/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.08, prob_loss:   1.71, total_loss:   3.90\n",
            " Total Loss  tf.Tensor(4.984795, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  123/125, lr:0.000005, giou_loss:   2.47, conf_loss:   0.18, prob_loss:   2.33, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(4.002139, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:  124/125, lr:0.000005, giou_loss:   1.91, conf_loss:   0.10, prob_loss:   1.99, total_loss:   4.00\n",
            " Total Loss  tf.Tensor(4.7002764, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:    0/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.07, prob_loss:   2.52, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(5.6638603, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:86 step:    1/125, lr:0.000005, giou_loss:   2.53, conf_loss:   0.17, prob_loss:   2.97, total_loss:   5.66\n",
            "\n",
            "\n",
            "giou_val_loss:   2.26, conf_val_loss:   0.64, prob_val_loss:   5.21, total_val_loss:   8.12\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.1071672, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    2/125, lr:0.000005, giou_loss:   1.90, conf_loss:   0.11, prob_loss:   2.10, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(5.0095625, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    3/125, lr:0.000005, giou_loss:   2.24, conf_loss:   0.09, prob_loss:   2.68, total_loss:   5.01\n",
            " Total Loss  tf.Tensor(5.3313704, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    4/125, lr:0.000005, giou_loss:   2.53, conf_loss:   0.23, prob_loss:   2.57, total_loss:   5.33\n",
            " Total Loss  tf.Tensor(4.710969, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    5/125, lr:0.000005, giou_loss:   2.25, conf_loss:   0.06, prob_loss:   2.40, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(4.823683, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    6/125, lr:0.000005, giou_loss:   2.38, conf_loss:   0.06, prob_loss:   2.39, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(4.4329944, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    7/125, lr:0.000005, giou_loss:   2.34, conf_loss:   0.01, prob_loss:   2.08, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(6.2159376, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    8/125, lr:0.000005, giou_loss:   3.17, conf_loss:   0.14, prob_loss:   2.91, total_loss:   6.22\n",
            " Total Loss  tf.Tensor(5.409812, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    9/125, lr:0.000005, giou_loss:   2.64, conf_loss:   0.15, prob_loss:   2.62, total_loss:   5.41\n",
            " Total Loss  tf.Tensor(3.631597, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   10/125, lr:0.000005, giou_loss:   1.69, conf_loss:   0.06, prob_loss:   1.88, total_loss:   3.63\n",
            " Total Loss  tf.Tensor(5.7251043, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   11/125, lr:0.000005, giou_loss:   2.51, conf_loss:   0.11, prob_loss:   3.10, total_loss:   5.73\n",
            " Total Loss  tf.Tensor(5.513431, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   12/125, lr:0.000005, giou_loss:   2.51, conf_loss:   0.17, prob_loss:   2.83, total_loss:   5.51\n",
            " Total Loss  tf.Tensor(4.7797356, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   13/125, lr:0.000005, giou_loss:   2.10, conf_loss:   0.49, prob_loss:   2.19, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(4.697461, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   14/125, lr:0.000005, giou_loss:   2.25, conf_loss:   0.13, prob_loss:   2.31, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(5.2362547, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   15/125, lr:0.000005, giou_loss:   2.40, conf_loss:   0.09, prob_loss:   2.75, total_loss:   5.24\n",
            " Total Loss  tf.Tensor(2.8022256, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   16/125, lr:0.000005, giou_loss:   1.28, conf_loss:   0.02, prob_loss:   1.50, total_loss:   2.80\n",
            " Total Loss  tf.Tensor(6.0484934, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   17/125, lr:0.000005, giou_loss:   2.93, conf_loss:   0.21, prob_loss:   2.91, total_loss:   6.05\n",
            " Total Loss  tf.Tensor(4.536377, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   18/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.07, prob_loss:   2.35, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(5.1124334, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   19/125, lr:0.000005, giou_loss:   2.35, conf_loss:   0.03, prob_loss:   2.72, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(5.6844654, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   20/125, lr:0.000005, giou_loss:   2.92, conf_loss:   0.04, prob_loss:   2.73, total_loss:   5.68\n",
            " Total Loss  tf.Tensor(4.6618204, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   21/125, lr:0.000005, giou_loss:   2.21, conf_loss:   0.20, prob_loss:   2.26, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.2730074, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   22/125, lr:0.000005, giou_loss:   2.06, conf_loss:   0.02, prob_loss:   2.20, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(4.165842, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   23/125, lr:0.000005, giou_loss:   2.14, conf_loss:   0.12, prob_loss:   1.91, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(3.5592446, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   24/125, lr:0.000005, giou_loss:   1.79, conf_loss:   0.14, prob_loss:   1.64, total_loss:   3.56\n",
            " Total Loss  tf.Tensor(5.442175, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   25/125, lr:0.000005, giou_loss:   2.60, conf_loss:   0.11, prob_loss:   2.73, total_loss:   5.44\n",
            " Total Loss  tf.Tensor(5.3881516, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   26/125, lr:0.000005, giou_loss:   2.48, conf_loss:   0.09, prob_loss:   2.82, total_loss:   5.39\n",
            " Total Loss  tf.Tensor(4.421941, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   27/125, lr:0.000005, giou_loss:   2.09, conf_loss:   0.02, prob_loss:   2.32, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(5.5074387, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   28/125, lr:0.000005, giou_loss:   2.59, conf_loss:   0.12, prob_loss:   2.80, total_loss:   5.51\n",
            " Total Loss  tf.Tensor(4.3365536, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   29/125, lr:0.000005, giou_loss:   2.14, conf_loss:   0.08, prob_loss:   2.11, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(5.4210052, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   30/125, lr:0.000005, giou_loss:   2.68, conf_loss:   0.09, prob_loss:   2.65, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(4.538563, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   31/125, lr:0.000005, giou_loss:   2.04, conf_loss:   0.02, prob_loss:   2.48, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(3.333373, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   32/125, lr:0.000005, giou_loss:   1.50, conf_loss:   0.07, prob_loss:   1.77, total_loss:   3.33\n",
            " Total Loss  tf.Tensor(5.0273695, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   33/125, lr:0.000005, giou_loss:   2.42, conf_loss:   0.08, prob_loss:   2.52, total_loss:   5.03\n",
            " Total Loss  tf.Tensor(4.4928136, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   34/125, lr:0.000005, giou_loss:   1.87, conf_loss:   0.07, prob_loss:   2.55, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(5.609334, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   35/125, lr:0.000005, giou_loss:   2.66, conf_loss:   0.24, prob_loss:   2.71, total_loss:   5.61\n",
            " Total Loss  tf.Tensor(3.9687328, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   36/125, lr:0.000005, giou_loss:   1.69, conf_loss:   0.02, prob_loss:   2.26, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(4.742869, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   37/125, lr:0.000005, giou_loss:   2.29, conf_loss:   0.07, prob_loss:   2.38, total_loss:   4.74\n",
            " Total Loss  tf.Tensor(4.435877, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   38/125, lr:0.000005, giou_loss:   2.23, conf_loss:   0.04, prob_loss:   2.17, total_loss:   4.44\n",
            " Total Loss  tf.Tensor(3.5353785, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   39/125, lr:0.000005, giou_loss:   1.85, conf_loss:   0.02, prob_loss:   1.66, total_loss:   3.54\n",
            " Total Loss  tf.Tensor(3.8021507, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   40/125, lr:0.000005, giou_loss:   1.80, conf_loss:   0.06, prob_loss:   1.94, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(3.939126, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   41/125, lr:0.000005, giou_loss:   1.84, conf_loss:   0.01, prob_loss:   2.09, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(4.5890017, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   42/125, lr:0.000005, giou_loss:   2.22, conf_loss:   0.02, prob_loss:   2.35, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(4.268895, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   43/125, lr:0.000005, giou_loss:   1.92, conf_loss:   0.03, prob_loss:   2.32, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(4.483503, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   44/125, lr:0.000005, giou_loss:   2.01, conf_loss:   0.15, prob_loss:   2.32, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(4.7097006, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   45/125, lr:0.000005, giou_loss:   2.07, conf_loss:   0.11, prob_loss:   2.53, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(3.3361413, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   46/125, lr:0.000005, giou_loss:   1.56, conf_loss:   0.09, prob_loss:   1.69, total_loss:   3.34\n",
            " Total Loss  tf.Tensor(3.8852768, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   47/125, lr:0.000005, giou_loss:   1.68, conf_loss:   0.08, prob_loss:   2.12, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(4.578721, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   48/125, lr:0.000005, giou_loss:   2.28, conf_loss:   0.06, prob_loss:   2.24, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.6677303, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   49/125, lr:0.000005, giou_loss:   1.72, conf_loss:   0.08, prob_loss:   1.87, total_loss:   3.67\n",
            " Total Loss  tf.Tensor(4.905076, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   50/125, lr:0.000005, giou_loss:   2.34, conf_loss:   0.03, prob_loss:   2.54, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(3.6829062, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   51/125, lr:0.000005, giou_loss:   1.78, conf_loss:   0.06, prob_loss:   1.84, total_loss:   3.68\n",
            " Total Loss  tf.Tensor(4.806117, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   52/125, lr:0.000005, giou_loss:   2.42, conf_loss:   0.25, prob_loss:   2.14, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(5.225633, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   53/125, lr:0.000005, giou_loss:   2.40, conf_loss:   0.23, prob_loss:   2.60, total_loss:   5.23\n",
            " Total Loss  tf.Tensor(4.6868486, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   54/125, lr:0.000005, giou_loss:   2.25, conf_loss:   0.12, prob_loss:   2.31, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(6.147368, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   55/125, lr:0.000005, giou_loss:   2.98, conf_loss:   0.19, prob_loss:   2.98, total_loss:   6.15\n",
            " Total Loss  tf.Tensor(6.023621, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   56/125, lr:0.000005, giou_loss:   2.78, conf_loss:   0.26, prob_loss:   2.99, total_loss:   6.02\n",
            " Total Loss  tf.Tensor(4.2798147, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   57/125, lr:0.000005, giou_loss:   1.94, conf_loss:   0.03, prob_loss:   2.31, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(4.5022526, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   58/125, lr:0.000005, giou_loss:   2.20, conf_loss:   0.14, prob_loss:   2.16, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.3474174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   59/125, lr:0.000005, giou_loss:   2.08, conf_loss:   0.04, prob_loss:   2.23, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(4.308486, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   60/125, lr:0.000005, giou_loss:   1.83, conf_loss:   0.25, prob_loss:   2.23, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(3.6755397, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   61/125, lr:0.000005, giou_loss:   2.03, conf_loss:   0.02, prob_loss:   1.63, total_loss:   3.68\n",
            " Total Loss  tf.Tensor(4.4253855, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   62/125, lr:0.000005, giou_loss:   1.98, conf_loss:   0.15, prob_loss:   2.30, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(4.8623104, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   63/125, lr:0.000005, giou_loss:   2.20, conf_loss:   0.20, prob_loss:   2.45, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(4.81726, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   64/125, lr:0.000005, giou_loss:   1.83, conf_loss:   0.15, prob_loss:   2.83, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(5.846792, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   65/125, lr:0.000005, giou_loss:   2.78, conf_loss:   0.05, prob_loss:   3.02, total_loss:   5.85\n",
            " Total Loss  tf.Tensor(5.4136267, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   66/125, lr:0.000005, giou_loss:   2.53, conf_loss:   0.07, prob_loss:   2.81, total_loss:   5.41\n",
            " Total Loss  tf.Tensor(4.978578, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   67/125, lr:0.000005, giou_loss:   2.34, conf_loss:   0.07, prob_loss:   2.57, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(4.9197636, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   68/125, lr:0.000005, giou_loss:   2.41, conf_loss:   0.04, prob_loss:   2.47, total_loss:   4.92\n",
            " Total Loss  tf.Tensor(5.573409, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   69/125, lr:0.000005, giou_loss:   2.54, conf_loss:   0.11, prob_loss:   2.93, total_loss:   5.57\n",
            " Total Loss  tf.Tensor(4.6230106, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   70/125, lr:0.000005, giou_loss:   2.18, conf_loss:   0.11, prob_loss:   2.33, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(3.5461683, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   71/125, lr:0.000005, giou_loss:   1.58, conf_loss:   0.04, prob_loss:   1.92, total_loss:   3.55\n",
            " Total Loss  tf.Tensor(4.4907517, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   72/125, lr:0.000005, giou_loss:   2.04, conf_loss:   0.24, prob_loss:   2.21, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(4.4544506, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   73/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.17, prob_loss:   2.18, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.43907, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   74/125, lr:0.000005, giou_loss:   2.26, conf_loss:   0.06, prob_loss:   2.11, total_loss:   4.44\n",
            " Total Loss  tf.Tensor(4.6480303, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   75/125, lr:0.000005, giou_loss:   2.10, conf_loss:   0.17, prob_loss:   2.38, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(4.582057, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   76/125, lr:0.000005, giou_loss:   2.37, conf_loss:   0.02, prob_loss:   2.19, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.270848, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   77/125, lr:0.000005, giou_loss:   1.99, conf_loss:   0.27, prob_loss:   2.01, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(3.6377935, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   78/125, lr:0.000005, giou_loss:   1.75, conf_loss:   0.03, prob_loss:   1.86, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(4.595128, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   79/125, lr:0.000005, giou_loss:   2.36, conf_loss:   0.11, prob_loss:   2.13, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(4.276492, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   80/125, lr:0.000005, giou_loss:   2.07, conf_loss:   0.03, prob_loss:   2.17, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(4.729068, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   81/125, lr:0.000005, giou_loss:   2.18, conf_loss:   0.08, prob_loss:   2.47, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(4.696566, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   82/125, lr:0.000005, giou_loss:   2.22, conf_loss:   0.06, prob_loss:   2.41, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(5.4377756, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   83/125, lr:0.000005, giou_loss:   2.56, conf_loss:   0.07, prob_loss:   2.81, total_loss:   5.44\n",
            " Total Loss  tf.Tensor(4.2543554, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   84/125, lr:0.000005, giou_loss:   1.97, conf_loss:   0.11, prob_loss:   2.18, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(3.5662982, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   85/125, lr:0.000005, giou_loss:   1.56, conf_loss:   0.03, prob_loss:   1.97, total_loss:   3.57\n",
            " Total Loss  tf.Tensor(3.4641128, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   86/125, lr:0.000005, giou_loss:   1.65, conf_loss:   0.04, prob_loss:   1.77, total_loss:   3.46\n",
            " Total Loss  tf.Tensor(4.230732, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   87/125, lr:0.000005, giou_loss:   1.99, conf_loss:   0.28, prob_loss:   1.96, total_loss:   4.23\n",
            " Total Loss  tf.Tensor(4.6880627, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   88/125, lr:0.000005, giou_loss:   2.22, conf_loss:   0.03, prob_loss:   2.44, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(3.4765568, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   89/125, lr:0.000005, giou_loss:   1.58, conf_loss:   0.06, prob_loss:   1.84, total_loss:   3.48\n",
            " Total Loss  tf.Tensor(3.972352, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   90/125, lr:0.000005, giou_loss:   1.74, conf_loss:   0.02, prob_loss:   2.21, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(4.036934, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   91/125, lr:0.000005, giou_loss:   1.82, conf_loss:   0.19, prob_loss:   2.03, total_loss:   4.04\n",
            " Total Loss  tf.Tensor(4.9983535, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   92/125, lr:0.000005, giou_loss:   2.49, conf_loss:   0.05, prob_loss:   2.46, total_loss:   5.00\n",
            " Total Loss  tf.Tensor(4.0083265, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   93/125, lr:0.000005, giou_loss:   1.91, conf_loss:   0.02, prob_loss:   2.08, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(5.127778, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   94/125, lr:0.000005, giou_loss:   2.44, conf_loss:   0.12, prob_loss:   2.57, total_loss:   5.13\n",
            " Total Loss  tf.Tensor(4.323682, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   95/125, lr:0.000005, giou_loss:   1.93, conf_loss:   0.12, prob_loss:   2.28, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(3.928749, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   96/125, lr:0.000005, giou_loss:   1.86, conf_loss:   0.03, prob_loss:   2.04, total_loss:   3.93\n",
            " Total Loss  tf.Tensor(4.870326, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   97/125, lr:0.000005, giou_loss:   2.24, conf_loss:   0.11, prob_loss:   2.52, total_loss:   4.87\n",
            " Total Loss  tf.Tensor(4.035977, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   98/125, lr:0.000005, giou_loss:   1.90, conf_loss:   0.06, prob_loss:   2.08, total_loss:   4.04\n",
            " Total Loss  tf.Tensor(4.0632405, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:   99/125, lr:0.000005, giou_loss:   1.79, conf_loss:   0.02, prob_loss:   2.25, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(5.7686243, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  100/125, lr:0.000005, giou_loss:   2.78, conf_loss:   0.21, prob_loss:   2.78, total_loss:   5.77\n",
            " Total Loss  tf.Tensor(5.549701, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  101/125, lr:0.000005, giou_loss:   2.41, conf_loss:   0.28, prob_loss:   2.86, total_loss:   5.55\n",
            " Total Loss  tf.Tensor(4.0516434, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  102/125, lr:0.000005, giou_loss:   1.75, conf_loss:   0.08, prob_loss:   2.22, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(4.424178, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  103/125, lr:0.000005, giou_loss:   2.07, conf_loss:   0.08, prob_loss:   2.27, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(6.0963945, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  104/125, lr:0.000005, giou_loss:   2.75, conf_loss:   0.67, prob_loss:   2.68, total_loss:   6.10\n",
            " Total Loss  tf.Tensor(4.2887297, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  105/125, lr:0.000005, giou_loss:   2.05, conf_loss:   0.17, prob_loss:   2.07, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(5.466568, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  106/125, lr:0.000005, giou_loss:   2.48, conf_loss:   0.06, prob_loss:   2.92, total_loss:   5.47\n",
            " Total Loss  tf.Tensor(4.292869, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  107/125, lr:0.000005, giou_loss:   1.86, conf_loss:   0.04, prob_loss:   2.39, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(3.639536, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  108/125, lr:0.000005, giou_loss:   1.65, conf_loss:   0.07, prob_loss:   1.92, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(4.7678094, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  109/125, lr:0.000005, giou_loss:   2.14, conf_loss:   0.06, prob_loss:   2.57, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(5.5971193, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  110/125, lr:0.000005, giou_loss:   2.57, conf_loss:   0.28, prob_loss:   2.75, total_loss:   5.60\n",
            " Total Loss  tf.Tensor(6.084796, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  111/125, lr:0.000005, giou_loss:   2.87, conf_loss:   0.08, prob_loss:   3.14, total_loss:   6.08\n",
            " Total Loss  tf.Tensor(3.799519, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  112/125, lr:0.000005, giou_loss:   1.74, conf_loss:   0.07, prob_loss:   1.99, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(4.3204894, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  113/125, lr:0.000005, giou_loss:   1.93, conf_loss:   0.33, prob_loss:   2.06, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(4.7013917, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  114/125, lr:0.000005, giou_loss:   2.00, conf_loss:   0.23, prob_loss:   2.47, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(5.1934266, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  115/125, lr:0.000005, giou_loss:   2.42, conf_loss:   0.05, prob_loss:   2.72, total_loss:   5.19\n",
            " Total Loss  tf.Tensor(5.5023966, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  116/125, lr:0.000005, giou_loss:   2.29, conf_loss:   0.09, prob_loss:   3.12, total_loss:   5.50\n",
            " Total Loss  tf.Tensor(4.159967, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  117/125, lr:0.000005, giou_loss:   1.81, conf_loss:   0.02, prob_loss:   2.33, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(4.5924706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  118/125, lr:0.000005, giou_loss:   1.68, conf_loss:   0.01, prob_loss:   2.89, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(4.6383576, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  119/125, lr:0.000005, giou_loss:   2.02, conf_loss:   0.42, prob_loss:   2.20, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.902332, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  120/125, lr:0.000005, giou_loss:   2.25, conf_loss:   0.07, prob_loss:   2.58, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(5.133889, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  121/125, lr:0.000005, giou_loss:   2.51, conf_loss:   0.10, prob_loss:   2.52, total_loss:   5.13\n",
            " Total Loss  tf.Tensor(4.305987, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  122/125, lr:0.000005, giou_loss:   2.00, conf_loss:   0.01, prob_loss:   2.29, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(3.8915677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  123/125, lr:0.000005, giou_loss:   2.03, conf_loss:   0.05, prob_loss:   1.81, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(4.9512954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:  124/125, lr:0.000005, giou_loss:   2.38, conf_loss:   0.04, prob_loss:   2.52, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(4.9314995, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    0/125, lr:0.000005, giou_loss:   2.52, conf_loss:   0.03, prob_loss:   2.38, total_loss:   4.93\n",
            " Total Loss  tf.Tensor(4.7659707, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:87 step:    1/125, lr:0.000005, giou_loss:   2.25, conf_loss:   0.13, prob_loss:   2.38, total_loss:   4.77\n",
            "\n",
            "\n",
            "giou_val_loss:   2.25, conf_val_loss:   0.64, prob_val_loss:   5.25, total_val_loss:   8.14\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(3.812357, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    2/125, lr:0.000005, giou_loss:   1.89, conf_loss:   0.03, prob_loss:   1.90, total_loss:   3.81\n",
            " Total Loss  tf.Tensor(4.822584, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    3/125, lr:0.000005, giou_loss:   2.33, conf_loss:   0.07, prob_loss:   2.42, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(4.556085, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    4/125, lr:0.000005, giou_loss:   2.29, conf_loss:   0.09, prob_loss:   2.17, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.2952576, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    5/125, lr:0.000005, giou_loss:   2.00, conf_loss:   0.08, prob_loss:   2.21, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.3048124, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    6/125, lr:0.000005, giou_loss:   1.95, conf_loss:   0.10, prob_loss:   2.26, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.877237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    7/125, lr:0.000005, giou_loss:   2.17, conf_loss:   0.03, prob_loss:   2.67, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(4.4984384, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    8/125, lr:0.000005, giou_loss:   2.01, conf_loss:   0.31, prob_loss:   2.18, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.754799, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    9/125, lr:0.000005, giou_loss:   2.29, conf_loss:   0.03, prob_loss:   2.44, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(3.920632, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   10/125, lr:0.000005, giou_loss:   1.75, conf_loss:   0.12, prob_loss:   2.06, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(5.034559, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   11/125, lr:0.000005, giou_loss:   2.44, conf_loss:   0.05, prob_loss:   2.54, total_loss:   5.03\n",
            " Total Loss  tf.Tensor(4.5067854, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   12/125, lr:0.000005, giou_loss:   2.04, conf_loss:   0.16, prob_loss:   2.30, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.7722106, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   13/125, lr:0.000005, giou_loss:   2.31, conf_loss:   0.02, prob_loss:   2.44, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(5.821507, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   14/125, lr:0.000005, giou_loss:   2.62, conf_loss:   0.07, prob_loss:   3.13, total_loss:   5.82\n",
            " Total Loss  tf.Tensor(5.800808, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   15/125, lr:0.000005, giou_loss:   2.74, conf_loss:   0.07, prob_loss:   2.98, total_loss:   5.80\n",
            " Total Loss  tf.Tensor(4.539562, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   16/125, lr:0.000005, giou_loss:   2.28, conf_loss:   0.02, prob_loss:   2.24, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(5.4244175, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   17/125, lr:0.000005, giou_loss:   2.67, conf_loss:   0.25, prob_loss:   2.50, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(4.300247, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   18/125, lr:0.000005, giou_loss:   2.03, conf_loss:   0.01, prob_loss:   2.26, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(3.7812734, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   19/125, lr:0.000005, giou_loss:   1.66, conf_loss:   0.05, prob_loss:   2.07, total_loss:   3.78\n",
            " Total Loss  tf.Tensor(4.949895, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   20/125, lr:0.000005, giou_loss:   2.52, conf_loss:   0.09, prob_loss:   2.34, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(4.340709, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   21/125, lr:0.000005, giou_loss:   2.04, conf_loss:   0.08, prob_loss:   2.22, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(4.2237654, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   22/125, lr:0.000005, giou_loss:   2.11, conf_loss:   0.03, prob_loss:   2.08, total_loss:   4.22\n",
            " Total Loss  tf.Tensor(4.9399815, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   23/125, lr:0.000005, giou_loss:   2.32, conf_loss:   0.08, prob_loss:   2.53, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(3.294001, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   24/125, lr:0.000005, giou_loss:   1.49, conf_loss:   0.03, prob_loss:   1.78, total_loss:   3.29\n",
            " Total Loss  tf.Tensor(3.7266617, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   25/125, lr:0.000004, giou_loss:   1.70, conf_loss:   0.08, prob_loss:   1.95, total_loss:   3.73\n",
            " Total Loss  tf.Tensor(5.0315866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   26/125, lr:0.000004, giou_loss:   2.36, conf_loss:   0.19, prob_loss:   2.48, total_loss:   5.03\n",
            " Total Loss  tf.Tensor(4.4203105, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   27/125, lr:0.000004, giou_loss:   2.07, conf_loss:   0.03, prob_loss:   2.32, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(4.1391125, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   28/125, lr:0.000004, giou_loss:   1.96, conf_loss:   0.19, prob_loss:   1.99, total_loss:   4.14\n",
            " Total Loss  tf.Tensor(5.44849, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   29/125, lr:0.000004, giou_loss:   2.56, conf_loss:   0.08, prob_loss:   2.81, total_loss:   5.45\n",
            " Total Loss  tf.Tensor(4.543624, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   30/125, lr:0.000004, giou_loss:   2.19, conf_loss:   0.04, prob_loss:   2.31, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(4.8817368, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   31/125, lr:0.000004, giou_loss:   1.87, conf_loss:   0.28, prob_loss:   2.72, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(5.6283617, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   32/125, lr:0.000004, giou_loss:   2.70, conf_loss:   0.19, prob_loss:   2.74, total_loss:   5.63\n",
            " Total Loss  tf.Tensor(4.485813, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   33/125, lr:0.000004, giou_loss:   2.06, conf_loss:   0.10, prob_loss:   2.33, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(4.851921, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   34/125, lr:0.000004, giou_loss:   2.25, conf_loss:   0.30, prob_loss:   2.30, total_loss:   4.85\n",
            " Total Loss  tf.Tensor(5.69886, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   35/125, lr:0.000004, giou_loss:   2.96, conf_loss:   0.18, prob_loss:   2.56, total_loss:   5.70\n",
            " Total Loss  tf.Tensor(5.543675, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   36/125, lr:0.000004, giou_loss:   2.67, conf_loss:   0.07, prob_loss:   2.80, total_loss:   5.54\n",
            " Total Loss  tf.Tensor(4.6016645, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   37/125, lr:0.000004, giou_loss:   1.97, conf_loss:   0.13, prob_loss:   2.50, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(5.653405, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   38/125, lr:0.000004, giou_loss:   2.72, conf_loss:   0.05, prob_loss:   2.89, total_loss:   5.65\n",
            " Total Loss  tf.Tensor(5.477252, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   39/125, lr:0.000004, giou_loss:   2.68, conf_loss:   0.19, prob_loss:   2.61, total_loss:   5.48\n",
            " Total Loss  tf.Tensor(4.0655737, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   40/125, lr:0.000004, giou_loss:   1.78, conf_loss:   0.21, prob_loss:   2.08, total_loss:   4.07\n",
            " Total Loss  tf.Tensor(5.3707504, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   41/125, lr:0.000004, giou_loss:   2.71, conf_loss:   0.21, prob_loss:   2.45, total_loss:   5.37\n",
            " Total Loss  tf.Tensor(3.909174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   42/125, lr:0.000004, giou_loss:   1.73, conf_loss:   0.04, prob_loss:   2.14, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.0988455, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   43/125, lr:0.000004, giou_loss:   1.94, conf_loss:   0.10, prob_loss:   2.05, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(5.1080914, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   44/125, lr:0.000004, giou_loss:   2.43, conf_loss:   0.21, prob_loss:   2.46, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(4.28518, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   45/125, lr:0.000004, giou_loss:   2.05, conf_loss:   0.10, prob_loss:   2.13, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.3234124, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   46/125, lr:0.000004, giou_loss:   2.04, conf_loss:   0.09, prob_loss:   2.19, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(5.1047797, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   47/125, lr:0.000004, giou_loss:   2.51, conf_loss:   0.02, prob_loss:   2.58, total_loss:   5.10\n",
            " Total Loss  tf.Tensor(5.418294, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   48/125, lr:0.000004, giou_loss:   2.40, conf_loss:   0.18, prob_loss:   2.84, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(4.4040318, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   49/125, lr:0.000004, giou_loss:   2.04, conf_loss:   0.35, prob_loss:   2.01, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(4.783147, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   50/125, lr:0.000004, giou_loss:   2.26, conf_loss:   0.06, prob_loss:   2.45, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(4.2102365, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   51/125, lr:0.000004, giou_loss:   2.00, conf_loss:   0.06, prob_loss:   2.15, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.609388, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   52/125, lr:0.000004, giou_loss:   2.03, conf_loss:   0.31, prob_loss:   2.27, total_loss:   4.61\n",
            " Total Loss  tf.Tensor(4.227643, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   53/125, lr:0.000004, giou_loss:   2.00, conf_loss:   0.02, prob_loss:   2.20, total_loss:   4.23\n",
            " Total Loss  tf.Tensor(5.6197786, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   54/125, lr:0.000004, giou_loss:   2.59, conf_loss:   0.24, prob_loss:   2.79, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(3.9235535, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   55/125, lr:0.000004, giou_loss:   1.71, conf_loss:   0.12, prob_loss:   2.09, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(4.9112606, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   56/125, lr:0.000004, giou_loss:   2.41, conf_loss:   0.04, prob_loss:   2.46, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(3.829771, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   57/125, lr:0.000004, giou_loss:   1.82, conf_loss:   0.13, prob_loss:   1.88, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(5.7460704, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   58/125, lr:0.000004, giou_loss:   2.55, conf_loss:   0.48, prob_loss:   2.72, total_loss:   5.75\n",
            " Total Loss  tf.Tensor(5.037901, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   59/125, lr:0.000004, giou_loss:   2.41, conf_loss:   0.19, prob_loss:   2.44, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(5.1891966, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   60/125, lr:0.000004, giou_loss:   2.46, conf_loss:   0.19, prob_loss:   2.54, total_loss:   5.19\n",
            " Total Loss  tf.Tensor(4.346602, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   61/125, lr:0.000004, giou_loss:   1.88, conf_loss:   0.05, prob_loss:   2.41, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(4.873908, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   62/125, lr:0.000004, giou_loss:   2.28, conf_loss:   0.04, prob_loss:   2.55, total_loss:   4.87\n",
            " Total Loss  tf.Tensor(4.5456667, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   63/125, lr:0.000004, giou_loss:   2.14, conf_loss:   0.04, prob_loss:   2.36, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(5.9032955, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   64/125, lr:0.000004, giou_loss:   2.74, conf_loss:   0.23, prob_loss:   2.94, total_loss:   5.90\n",
            " Total Loss  tf.Tensor(4.1360855, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   65/125, lr:0.000004, giou_loss:   2.10, conf_loss:   0.01, prob_loss:   2.02, total_loss:   4.14\n",
            " Total Loss  tf.Tensor(4.4153805, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   66/125, lr:0.000004, giou_loss:   2.12, conf_loss:   0.04, prob_loss:   2.26, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(5.4157515, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   67/125, lr:0.000004, giou_loss:   1.96, conf_loss:   0.03, prob_loss:   3.43, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(4.969061, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   68/125, lr:0.000004, giou_loss:   2.26, conf_loss:   0.04, prob_loss:   2.67, total_loss:   4.97\n",
            " Total Loss  tf.Tensor(4.8897843, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   69/125, lr:0.000004, giou_loss:   2.37, conf_loss:   0.13, prob_loss:   2.39, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(4.902033, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   70/125, lr:0.000004, giou_loss:   2.44, conf_loss:   0.20, prob_loss:   2.26, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(3.7671947, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   71/125, lr:0.000004, giou_loss:   1.77, conf_loss:   0.01, prob_loss:   1.99, total_loss:   3.77\n",
            " Total Loss  tf.Tensor(4.256917, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   72/125, lr:0.000004, giou_loss:   2.02, conf_loss:   0.03, prob_loss:   2.20, total_loss:   4.26\n",
            " Total Loss  tf.Tensor(5.3850327, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   73/125, lr:0.000004, giou_loss:   2.69, conf_loss:   0.14, prob_loss:   2.55, total_loss:   5.39\n",
            " Total Loss  tf.Tensor(3.3222108, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   74/125, lr:0.000004, giou_loss:   1.58, conf_loss:   0.03, prob_loss:   1.71, total_loss:   3.32\n",
            " Total Loss  tf.Tensor(3.6884027, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   75/125, lr:0.000004, giou_loss:   1.49, conf_loss:   0.02, prob_loss:   2.18, total_loss:   3.69\n",
            " Total Loss  tf.Tensor(3.845674, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   76/125, lr:0.000004, giou_loss:   1.76, conf_loss:   0.01, prob_loss:   2.07, total_loss:   3.85\n",
            " Total Loss  tf.Tensor(4.090532, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   77/125, lr:0.000004, giou_loss:   1.98, conf_loss:   0.07, prob_loss:   2.03, total_loss:   4.09\n",
            " Total Loss  tf.Tensor(3.9782887, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   78/125, lr:0.000004, giou_loss:   1.48, conf_loss:   0.32, prob_loss:   2.18, total_loss:   3.98\n",
            " Total Loss  tf.Tensor(3.9423752, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   79/125, lr:0.000004, giou_loss:   1.87, conf_loss:   0.09, prob_loss:   1.99, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(5.261136, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   80/125, lr:0.000004, giou_loss:   2.49, conf_loss:   0.08, prob_loss:   2.69, total_loss:   5.26\n",
            " Total Loss  tf.Tensor(4.9203024, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   81/125, lr:0.000004, giou_loss:   2.41, conf_loss:   0.20, prob_loss:   2.31, total_loss:   4.92\n",
            " Total Loss  tf.Tensor(4.5104675, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   82/125, lr:0.000004, giou_loss:   2.11, conf_loss:   0.35, prob_loss:   2.05, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.121921, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   83/125, lr:0.000004, giou_loss:   1.76, conf_loss:   0.03, prob_loss:   2.33, total_loss:   4.12\n",
            " Total Loss  tf.Tensor(4.888666, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   84/125, lr:0.000004, giou_loss:   2.31, conf_loss:   0.04, prob_loss:   2.54, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(5.6019163, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   85/125, lr:0.000004, giou_loss:   2.73, conf_loss:   0.32, prob_loss:   2.55, total_loss:   5.60\n",
            " Total Loss  tf.Tensor(2.7282138, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   86/125, lr:0.000004, giou_loss:   1.18, conf_loss:   0.04, prob_loss:   1.51, total_loss:   2.73\n",
            " Total Loss  tf.Tensor(4.6580696, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   87/125, lr:0.000004, giou_loss:   2.19, conf_loss:   0.02, prob_loss:   2.44, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.481016, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   88/125, lr:0.000004, giou_loss:   2.18, conf_loss:   0.04, prob_loss:   2.26, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(4.2936907, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   89/125, lr:0.000004, giou_loss:   1.87, conf_loss:   0.14, prob_loss:   2.28, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.0449486, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   90/125, lr:0.000004, giou_loss:   1.79, conf_loss:   0.16, prob_loss:   2.09, total_loss:   4.04\n",
            " Total Loss  tf.Tensor(4.2993197, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   91/125, lr:0.000004, giou_loss:   2.03, conf_loss:   0.07, prob_loss:   2.20, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(5.085326, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   92/125, lr:0.000004, giou_loss:   2.38, conf_loss:   0.23, prob_loss:   2.47, total_loss:   5.09\n",
            " Total Loss  tf.Tensor(3.7984862, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   93/125, lr:0.000004, giou_loss:   1.64, conf_loss:   0.07, prob_loss:   2.08, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(5.039153, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   94/125, lr:0.000004, giou_loss:   2.18, conf_loss:   0.07, prob_loss:   2.78, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(6.034696, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   95/125, lr:0.000004, giou_loss:   2.35, conf_loss:   0.27, prob_loss:   3.41, total_loss:   6.03\n",
            " Total Loss  tf.Tensor(5.1409426, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   96/125, lr:0.000004, giou_loss:   2.44, conf_loss:   0.20, prob_loss:   2.50, total_loss:   5.14\n",
            " Total Loss  tf.Tensor(3.759432, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   97/125, lr:0.000004, giou_loss:   1.61, conf_loss:   0.17, prob_loss:   1.98, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(5.7064342, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   98/125, lr:0.000004, giou_loss:   2.62, conf_loss:   0.40, prob_loss:   2.69, total_loss:   5.71\n",
            " Total Loss  tf.Tensor(3.8048086, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:   99/125, lr:0.000004, giou_loss:   1.75, conf_loss:   0.07, prob_loss:   1.98, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(4.434425, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  100/125, lr:0.000004, giou_loss:   2.01, conf_loss:   0.13, prob_loss:   2.29, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(4.146508, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  101/125, lr:0.000004, giou_loss:   1.95, conf_loss:   0.04, prob_loss:   2.16, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(5.1741066, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  102/125, lr:0.000004, giou_loss:   2.34, conf_loss:   0.03, prob_loss:   2.81, total_loss:   5.17\n",
            " Total Loss  tf.Tensor(4.490586, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  103/125, lr:0.000004, giou_loss:   2.11, conf_loss:   0.04, prob_loss:   2.34, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(5.438739, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  104/125, lr:0.000004, giou_loss:   2.61, conf_loss:   0.09, prob_loss:   2.73, total_loss:   5.44\n",
            " Total Loss  tf.Tensor(3.7782998, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  105/125, lr:0.000004, giou_loss:   1.81, conf_loss:   0.06, prob_loss:   1.91, total_loss:   3.78\n",
            " Total Loss  tf.Tensor(5.2127867, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  106/125, lr:0.000004, giou_loss:   2.40, conf_loss:   0.02, prob_loss:   2.79, total_loss:   5.21\n",
            " Total Loss  tf.Tensor(4.4602823, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  107/125, lr:0.000004, giou_loss:   2.28, conf_loss:   0.11, prob_loss:   2.07, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(4.3119655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  108/125, lr:0.000004, giou_loss:   2.15, conf_loss:   0.02, prob_loss:   2.14, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.385631, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  109/125, lr:0.000004, giou_loss:   1.98, conf_loss:   0.18, prob_loss:   2.22, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(4.315109, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  110/125, lr:0.000004, giou_loss:   2.13, conf_loss:   0.15, prob_loss:   2.03, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(4.089802, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  111/125, lr:0.000004, giou_loss:   1.83, conf_loss:   0.09, prob_loss:   2.17, total_loss:   4.09\n",
            " Total Loss  tf.Tensor(4.769254, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  112/125, lr:0.000004, giou_loss:   2.13, conf_loss:   0.06, prob_loss:   2.57, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(5.1644964, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  113/125, lr:0.000004, giou_loss:   2.23, conf_loss:   0.17, prob_loss:   2.77, total_loss:   5.16\n",
            " Total Loss  tf.Tensor(4.8948274, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  114/125, lr:0.000004, giou_loss:   2.13, conf_loss:   0.03, prob_loss:   2.74, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(5.222193, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  115/125, lr:0.000004, giou_loss:   2.47, conf_loss:   0.04, prob_loss:   2.70, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(4.5003576, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  116/125, lr:0.000004, giou_loss:   2.11, conf_loss:   0.12, prob_loss:   2.27, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.858182, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  117/125, lr:0.000004, giou_loss:   2.14, conf_loss:   0.11, prob_loss:   2.61, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(4.287349, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  118/125, lr:0.000004, giou_loss:   1.83, conf_loss:   0.10, prob_loss:   2.36, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(5.3467345, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  119/125, lr:0.000004, giou_loss:   2.30, conf_loss:   0.18, prob_loss:   2.86, total_loss:   5.35\n",
            " Total Loss  tf.Tensor(3.9039178, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  120/125, lr:0.000004, giou_loss:   1.85, conf_loss:   0.07, prob_loss:   1.98, total_loss:   3.90\n",
            " Total Loss  tf.Tensor(5.736918, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  121/125, lr:0.000004, giou_loss:   2.80, conf_loss:   0.05, prob_loss:   2.88, total_loss:   5.74\n",
            " Total Loss  tf.Tensor(4.1231594, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  122/125, lr:0.000004, giou_loss:   1.86, conf_loss:   0.05, prob_loss:   2.21, total_loss:   4.12\n",
            " Total Loss  tf.Tensor(4.9623876, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  123/125, lr:0.000004, giou_loss:   2.55, conf_loss:   0.12, prob_loss:   2.29, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(4.568694, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:  124/125, lr:0.000004, giou_loss:   2.06, conf_loss:   0.22, prob_loss:   2.29, total_loss:   4.57\n",
            " Total Loss  tf.Tensor(4.834278, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    0/125, lr:0.000004, giou_loss:   2.20, conf_loss:   0.05, prob_loss:   2.59, total_loss:   4.83\n",
            " Total Loss  tf.Tensor(4.680229, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:88 step:    1/125, lr:0.000004, giou_loss:   2.14, conf_loss:   0.10, prob_loss:   2.45, total_loss:   4.68\n",
            "\n",
            "\n",
            "giou_val_loss:   2.29, conf_val_loss:   0.62, prob_val_loss:   5.33, total_val_loss:   8.23\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.6062255, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    2/125, lr:0.000004, giou_loss:   2.21, conf_loss:   0.06, prob_loss:   2.34, total_loss:   4.61\n",
            " Total Loss  tf.Tensor(4.1962442, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    3/125, lr:0.000004, giou_loss:   2.00, conf_loss:   0.05, prob_loss:   2.14, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(5.339404, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    4/125, lr:0.000004, giou_loss:   2.49, conf_loss:   0.07, prob_loss:   2.78, total_loss:   5.34\n",
            " Total Loss  tf.Tensor(5.3727636, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    5/125, lr:0.000004, giou_loss:   2.29, conf_loss:   0.05, prob_loss:   3.03, total_loss:   5.37\n",
            " Total Loss  tf.Tensor(3.6390495, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    6/125, lr:0.000004, giou_loss:   1.90, conf_loss:   0.06, prob_loss:   1.68, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(5.2186556, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    7/125, lr:0.000004, giou_loss:   2.45, conf_loss:   0.08, prob_loss:   2.69, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(4.85795, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    8/125, lr:0.000004, giou_loss:   2.15, conf_loss:   0.16, prob_loss:   2.55, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(4.5072346, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    9/125, lr:0.000004, giou_loss:   2.06, conf_loss:   0.02, prob_loss:   2.43, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.1650524, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   10/125, lr:0.000004, giou_loss:   1.82, conf_loss:   0.49, prob_loss:   1.85, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.2710824, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   11/125, lr:0.000004, giou_loss:   1.80, conf_loss:   0.43, prob_loss:   2.04, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(6.304924, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   12/125, lr:0.000004, giou_loss:   3.00, conf_loss:   0.34, prob_loss:   2.97, total_loss:   6.30\n",
            " Total Loss  tf.Tensor(5.0973496, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   13/125, lr:0.000004, giou_loss:   2.35, conf_loss:   0.10, prob_loss:   2.65, total_loss:   5.10\n",
            " Total Loss  tf.Tensor(5.463364, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   14/125, lr:0.000004, giou_loss:   2.36, conf_loss:   0.18, prob_loss:   2.92, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(4.6624126, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   15/125, lr:0.000004, giou_loss:   2.26, conf_loss:   0.11, prob_loss:   2.29, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.5480514, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   16/125, lr:0.000004, giou_loss:   2.15, conf_loss:   0.02, prob_loss:   2.39, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(3.6084547, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   17/125, lr:0.000004, giou_loss:   1.76, conf_loss:   0.11, prob_loss:   1.74, total_loss:   3.61\n",
            " Total Loss  tf.Tensor(5.5974035, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   18/125, lr:0.000004, giou_loss:   2.91, conf_loss:   0.10, prob_loss:   2.58, total_loss:   5.60\n",
            " Total Loss  tf.Tensor(3.9930992, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   19/125, lr:0.000004, giou_loss:   1.87, conf_loss:   0.07, prob_loss:   2.05, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(5.7946568, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   20/125, lr:0.000004, giou_loss:   2.71, conf_loss:   0.07, prob_loss:   3.01, total_loss:   5.79\n",
            " Total Loss  tf.Tensor(3.323167, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   21/125, lr:0.000004, giou_loss:   1.42, conf_loss:   0.05, prob_loss:   1.85, total_loss:   3.32\n",
            " Total Loss  tf.Tensor(4.363553, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   22/125, lr:0.000004, giou_loss:   2.04, conf_loss:   0.13, prob_loss:   2.19, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.7060604, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   23/125, lr:0.000004, giou_loss:   2.26, conf_loss:   0.14, prob_loss:   2.31, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(5.059005, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   24/125, lr:0.000004, giou_loss:   2.62, conf_loss:   0.08, prob_loss:   2.37, total_loss:   5.06\n",
            " Total Loss  tf.Tensor(4.3665743, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   25/125, lr:0.000004, giou_loss:   1.95, conf_loss:   0.08, prob_loss:   2.33, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(5.5327816, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   26/125, lr:0.000004, giou_loss:   2.57, conf_loss:   0.06, prob_loss:   2.90, total_loss:   5.53\n",
            " Total Loss  tf.Tensor(4.0968304, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   27/125, lr:0.000004, giou_loss:   1.99, conf_loss:   0.07, prob_loss:   2.03, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(8.0389, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   28/125, lr:0.000004, giou_loss:   2.54, conf_loss:   2.46, prob_loss:   3.03, total_loss:   8.04\n",
            " Total Loss  tf.Tensor(4.95195, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   29/125, lr:0.000004, giou_loss:   2.30, conf_loss:   0.05, prob_loss:   2.60, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(4.3367953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   30/125, lr:0.000004, giou_loss:   1.92, conf_loss:   0.02, prob_loss:   2.39, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(3.5250187, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   31/125, lr:0.000004, giou_loss:   1.74, conf_loss:   0.06, prob_loss:   1.72, total_loss:   3.53\n",
            " Total Loss  tf.Tensor(5.0589542, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   32/125, lr:0.000004, giou_loss:   2.43, conf_loss:   0.10, prob_loss:   2.53, total_loss:   5.06\n",
            " Total Loss  tf.Tensor(5.0440054, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   33/125, lr:0.000004, giou_loss:   2.31, conf_loss:   0.27, prob_loss:   2.47, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(5.147674, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   34/125, lr:0.000004, giou_loss:   2.42, conf_loss:   0.18, prob_loss:   2.55, total_loss:   5.15\n",
            " Total Loss  tf.Tensor(3.9721653, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   35/125, lr:0.000004, giou_loss:   2.01, conf_loss:   0.03, prob_loss:   1.93, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(3.755508, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   36/125, lr:0.000004, giou_loss:   1.72, conf_loss:   0.07, prob_loss:   1.97, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(4.8836794, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   37/125, lr:0.000004, giou_loss:   2.32, conf_loss:   0.16, prob_loss:   2.40, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(3.6380749, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   38/125, lr:0.000004, giou_loss:   1.68, conf_loss:   0.02, prob_loss:   1.93, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(4.2545652, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   39/125, lr:0.000004, giou_loss:   2.11, conf_loss:   0.04, prob_loss:   2.11, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(6.1543427, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   40/125, lr:0.000004, giou_loss:   2.89, conf_loss:   0.53, prob_loss:   2.74, total_loss:   6.15\n",
            " Total Loss  tf.Tensor(4.0672245, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   41/125, lr:0.000004, giou_loss:   1.89, conf_loss:   0.05, prob_loss:   2.13, total_loss:   4.07\n",
            " Total Loss  tf.Tensor(5.2636886, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   42/125, lr:0.000004, giou_loss:   2.52, conf_loss:   0.14, prob_loss:   2.61, total_loss:   5.26\n",
            " Total Loss  tf.Tensor(4.3779573, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   43/125, lr:0.000004, giou_loss:   1.83, conf_loss:   0.10, prob_loss:   2.45, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(5.0934157, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   44/125, lr:0.000004, giou_loss:   2.29, conf_loss:   0.05, prob_loss:   2.76, total_loss:   5.09\n",
            " Total Loss  tf.Tensor(4.7066336, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   45/125, lr:0.000004, giou_loss:   2.19, conf_loss:   0.03, prob_loss:   2.49, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(5.4982595, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   46/125, lr:0.000004, giou_loss:   2.82, conf_loss:   0.16, prob_loss:   2.52, total_loss:   5.50\n",
            " Total Loss  tf.Tensor(2.7163734, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   47/125, lr:0.000004, giou_loss:   1.16, conf_loss:   0.01, prob_loss:   1.55, total_loss:   2.72\n",
            " Total Loss  tf.Tensor(3.8437138, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   48/125, lr:0.000004, giou_loss:   1.79, conf_loss:   0.06, prob_loss:   1.99, total_loss:   3.84\n",
            " Total Loss  tf.Tensor(4.785721, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   49/125, lr:0.000004, giou_loss:   2.22, conf_loss:   0.15, prob_loss:   2.41, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(5.1085095, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   50/125, lr:0.000004, giou_loss:   2.48, conf_loss:   0.07, prob_loss:   2.55, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(4.145318, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   51/125, lr:0.000004, giou_loss:   1.93, conf_loss:   0.04, prob_loss:   2.18, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(4.879407, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   52/125, lr:0.000004, giou_loss:   2.41, conf_loss:   0.06, prob_loss:   2.41, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(4.80594, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   53/125, lr:0.000004, giou_loss:   2.13, conf_loss:   0.10, prob_loss:   2.57, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(3.6038823, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   54/125, lr:0.000004, giou_loss:   1.77, conf_loss:   0.07, prob_loss:   1.77, total_loss:   3.60\n",
            " Total Loss  tf.Tensor(3.8322277, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   55/125, lr:0.000004, giou_loss:   1.92, conf_loss:   0.11, prob_loss:   1.80, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(5.2483764, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   56/125, lr:0.000004, giou_loss:   2.40, conf_loss:   0.32, prob_loss:   2.53, total_loss:   5.25\n",
            " Total Loss  tf.Tensor(3.916471, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   57/125, lr:0.000004, giou_loss:   1.75, conf_loss:   0.08, prob_loss:   2.09, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(5.4766016, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   58/125, lr:0.000004, giou_loss:   2.69, conf_loss:   0.21, prob_loss:   2.57, total_loss:   5.48\n",
            " Total Loss  tf.Tensor(4.319585, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   59/125, lr:0.000004, giou_loss:   2.05, conf_loss:   0.21, prob_loss:   2.06, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(3.9601421, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   60/125, lr:0.000004, giou_loss:   1.73, conf_loss:   0.18, prob_loss:   2.05, total_loss:   3.96\n",
            " Total Loss  tf.Tensor(5.2647424, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   61/125, lr:0.000004, giou_loss:   2.46, conf_loss:   0.40, prob_loss:   2.41, total_loss:   5.26\n",
            " Total Loss  tf.Tensor(3.5358732, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   62/125, lr:0.000004, giou_loss:   1.80, conf_loss:   0.08, prob_loss:   1.66, total_loss:   3.54\n",
            " Total Loss  tf.Tensor(4.4580374, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   63/125, lr:0.000004, giou_loss:   2.27, conf_loss:   0.14, prob_loss:   2.05, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(3.2645195, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   64/125, lr:0.000004, giou_loss:   1.51, conf_loss:   0.04, prob_loss:   1.72, total_loss:   3.26\n",
            " Total Loss  tf.Tensor(5.573304, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   65/125, lr:0.000004, giou_loss:   2.76, conf_loss:   0.12, prob_loss:   2.70, total_loss:   5.57\n",
            " Total Loss  tf.Tensor(4.344286, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   66/125, lr:0.000004, giou_loss:   2.28, conf_loss:   0.06, prob_loss:   2.01, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(4.169458, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   67/125, lr:0.000004, giou_loss:   2.05, conf_loss:   0.04, prob_loss:   2.08, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(3.5693338, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   68/125, lr:0.000004, giou_loss:   1.72, conf_loss:   0.04, prob_loss:   1.80, total_loss:   3.57\n",
            " Total Loss  tf.Tensor(4.639571, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   69/125, lr:0.000004, giou_loss:   2.33, conf_loss:   0.03, prob_loss:   2.28, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(6.095595, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   70/125, lr:0.000004, giou_loss:   2.64, conf_loss:   0.18, prob_loss:   3.28, total_loss:   6.10\n",
            " Total Loss  tf.Tensor(4.6244, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   71/125, lr:0.000004, giou_loss:   2.09, conf_loss:   0.06, prob_loss:   2.47, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(5.401436, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   72/125, lr:0.000004, giou_loss:   2.70, conf_loss:   0.40, prob_loss:   2.29, total_loss:   5.40\n",
            " Total Loss  tf.Tensor(4.5981874, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   73/125, lr:0.000004, giou_loss:   1.89, conf_loss:   0.03, prob_loss:   2.68, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(3.5938442, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   74/125, lr:0.000004, giou_loss:   1.77, conf_loss:   0.06, prob_loss:   1.76, total_loss:   3.59\n",
            " Total Loss  tf.Tensor(5.541777, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   75/125, lr:0.000004, giou_loss:   2.68, conf_loss:   0.03, prob_loss:   2.84, total_loss:   5.54\n",
            " Total Loss  tf.Tensor(4.0328565, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   76/125, lr:0.000004, giou_loss:   2.00, conf_loss:   0.02, prob_loss:   2.00, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(4.5459538, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   77/125, lr:0.000004, giou_loss:   2.20, conf_loss:   0.10, prob_loss:   2.25, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(4.663513, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   78/125, lr:0.000004, giou_loss:   2.20, conf_loss:   0.07, prob_loss:   2.39, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.3317003, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   79/125, lr:0.000004, giou_loss:   2.13, conf_loss:   0.09, prob_loss:   2.11, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(5.9036922, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   80/125, lr:0.000004, giou_loss:   2.79, conf_loss:   0.09, prob_loss:   3.02, total_loss:   5.90\n",
            " Total Loss  tf.Tensor(4.221014, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   81/125, lr:0.000004, giou_loss:   2.03, conf_loss:   0.03, prob_loss:   2.16, total_loss:   4.22\n",
            " Total Loss  tf.Tensor(5.1154275, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   82/125, lr:0.000004, giou_loss:   2.43, conf_loss:   0.10, prob_loss:   2.58, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(4.7737026, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   83/125, lr:0.000004, giou_loss:   2.20, conf_loss:   0.03, prob_loss:   2.55, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(4.745278, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   84/125, lr:0.000004, giou_loss:   2.22, conf_loss:   0.31, prob_loss:   2.21, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(4.5122843, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   85/125, lr:0.000004, giou_loss:   2.01, conf_loss:   0.03, prob_loss:   2.47, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(5.254881, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   86/125, lr:0.000004, giou_loss:   2.49, conf_loss:   0.09, prob_loss:   2.67, total_loss:   5.25\n",
            " Total Loss  tf.Tensor(4.789006, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   87/125, lr:0.000004, giou_loss:   2.24, conf_loss:   0.07, prob_loss:   2.48, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(5.492156, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   88/125, lr:0.000004, giou_loss:   2.54, conf_loss:   0.07, prob_loss:   2.89, total_loss:   5.49\n",
            " Total Loss  tf.Tensor(4.839142, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   89/125, lr:0.000004, giou_loss:   2.11, conf_loss:   0.12, prob_loss:   2.61, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.0009923, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   90/125, lr:0.000004, giou_loss:   1.96, conf_loss:   0.11, prob_loss:   1.93, total_loss:   4.00\n",
            " Total Loss  tf.Tensor(4.636741, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   91/125, lr:0.000004, giou_loss:   2.01, conf_loss:   0.06, prob_loss:   2.57, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(3.0555086, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   92/125, lr:0.000004, giou_loss:   1.43, conf_loss:   0.02, prob_loss:   1.60, total_loss:   3.06\n",
            " Total Loss  tf.Tensor(5.3040476, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   93/125, lr:0.000004, giou_loss:   2.30, conf_loss:   0.03, prob_loss:   2.97, total_loss:   5.30\n",
            " Total Loss  tf.Tensor(3.5786438, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   94/125, lr:0.000004, giou_loss:   1.59, conf_loss:   0.02, prob_loss:   1.97, total_loss:   3.58\n",
            " Total Loss  tf.Tensor(4.1345167, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   95/125, lr:0.000004, giou_loss:   1.90, conf_loss:   0.09, prob_loss:   2.15, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(6.094817, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   96/125, lr:0.000004, giou_loss:   2.84, conf_loss:   0.13, prob_loss:   3.13, total_loss:   6.09\n",
            " Total Loss  tf.Tensor(4.31314, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   97/125, lr:0.000004, giou_loss:   1.98, conf_loss:   0.07, prob_loss:   2.26, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(3.7602558, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   98/125, lr:0.000004, giou_loss:   1.78, conf_loss:   0.33, prob_loss:   1.65, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(5.233971, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:   99/125, lr:0.000004, giou_loss:   2.34, conf_loss:   0.07, prob_loss:   2.83, total_loss:   5.23\n",
            " Total Loss  tf.Tensor(5.791988, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  100/125, lr:0.000004, giou_loss:   2.82, conf_loss:   0.28, prob_loss:   2.69, total_loss:   5.79\n",
            " Total Loss  tf.Tensor(4.4511557, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  101/125, lr:0.000004, giou_loss:   2.18, conf_loss:   0.10, prob_loss:   2.17, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(5.666127, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  102/125, lr:0.000004, giou_loss:   2.68, conf_loss:   0.25, prob_loss:   2.74, total_loss:   5.67\n",
            " Total Loss  tf.Tensor(4.060606, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  103/125, lr:0.000004, giou_loss:   1.87, conf_loss:   0.08, prob_loss:   2.11, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(4.438501, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  104/125, lr:0.000004, giou_loss:   1.94, conf_loss:   0.04, prob_loss:   2.47, total_loss:   4.44\n",
            " Total Loss  tf.Tensor(5.6206574, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  105/125, lr:0.000004, giou_loss:   2.65, conf_loss:   0.28, prob_loss:   2.69, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(3.596551, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  106/125, lr:0.000004, giou_loss:   1.72, conf_loss:   0.02, prob_loss:   1.86, total_loss:   3.60\n",
            " Total Loss  tf.Tensor(4.5779896, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  107/125, lr:0.000004, giou_loss:   2.03, conf_loss:   0.15, prob_loss:   2.40, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.3651237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  108/125, lr:0.000004, giou_loss:   2.08, conf_loss:   0.03, prob_loss:   2.26, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(4.5072536, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  109/125, lr:0.000004, giou_loss:   2.12, conf_loss:   0.02, prob_loss:   2.36, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(5.5293303, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  110/125, lr:0.000004, giou_loss:   2.66, conf_loss:   0.15, prob_loss:   2.71, total_loss:   5.53\n",
            " Total Loss  tf.Tensor(3.9525325, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  111/125, lr:0.000004, giou_loss:   1.61, conf_loss:   0.14, prob_loss:   2.20, total_loss:   3.95\n",
            " Total Loss  tf.Tensor(4.1093426, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  112/125, lr:0.000004, giou_loss:   1.80, conf_loss:   0.02, prob_loss:   2.29, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(5.4670963, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  113/125, lr:0.000004, giou_loss:   2.62, conf_loss:   0.05, prob_loss:   2.80, total_loss:   5.47\n",
            " Total Loss  tf.Tensor(3.954677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  114/125, lr:0.000004, giou_loss:   1.96, conf_loss:   0.12, prob_loss:   1.87, total_loss:   3.95\n",
            " Total Loss  tf.Tensor(4.8992434, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  115/125, lr:0.000004, giou_loss:   2.33, conf_loss:   0.03, prob_loss:   2.54, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(3.9135327, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  116/125, lr:0.000004, giou_loss:   1.82, conf_loss:   0.11, prob_loss:   1.98, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.1342645, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  117/125, lr:0.000004, giou_loss:   1.88, conf_loss:   0.10, prob_loss:   2.15, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(4.0696507, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  118/125, lr:0.000004, giou_loss:   1.77, conf_loss:   0.05, prob_loss:   2.25, total_loss:   4.07\n",
            " Total Loss  tf.Tensor(4.1922245, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  119/125, lr:0.000004, giou_loss:   1.95, conf_loss:   0.08, prob_loss:   2.16, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(4.419463, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  120/125, lr:0.000004, giou_loss:   2.06, conf_loss:   0.02, prob_loss:   2.33, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(5.0537567, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  121/125, lr:0.000004, giou_loss:   2.54, conf_loss:   0.09, prob_loss:   2.42, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(4.6662493, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  122/125, lr:0.000004, giou_loss:   2.20, conf_loss:   0.33, prob_loss:   2.13, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.4645796, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  123/125, lr:0.000004, giou_loss:   2.78, conf_loss:   0.10, prob_loss:   2.58, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(4.4728746, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:  124/125, lr:0.000004, giou_loss:   1.97, conf_loss:   0.16, prob_loss:   2.34, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(3.7517686, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    0/125, lr:0.000004, giou_loss:   1.88, conf_loss:   0.06, prob_loss:   1.81, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(5.1004305, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:89 step:    1/125, lr:0.000004, giou_loss:   2.52, conf_loss:   0.05, prob_loss:   2.53, total_loss:   5.10\n",
            "\n",
            "\n",
            "giou_val_loss:   2.26, conf_val_loss:   0.61, prob_val_loss:   5.39, total_val_loss:   8.25\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.5033875, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    2/125, lr:0.000004, giou_loss:   1.98, conf_loss:   0.04, prob_loss:   2.48, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(5.324472, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    3/125, lr:0.000004, giou_loss:   2.37, conf_loss:   0.21, prob_loss:   2.75, total_loss:   5.32\n",
            " Total Loss  tf.Tensor(3.6245558, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    4/125, lr:0.000004, giou_loss:   1.78, conf_loss:   0.06, prob_loss:   1.78, total_loss:   3.62\n",
            " Total Loss  tf.Tensor(4.842104, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    5/125, lr:0.000004, giou_loss:   2.46, conf_loss:   0.02, prob_loss:   2.36, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(3.9109106, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    6/125, lr:0.000003, giou_loss:   1.78, conf_loss:   0.03, prob_loss:   2.10, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.379265, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    7/125, lr:0.000003, giou_loss:   1.91, conf_loss:   0.05, prob_loss:   2.42, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(5.5804677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    8/125, lr:0.000003, giou_loss:   2.32, conf_loss:   0.40, prob_loss:   2.86, total_loss:   5.58\n",
            " Total Loss  tf.Tensor(5.0774374, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    9/125, lr:0.000003, giou_loss:   2.51, conf_loss:   0.04, prob_loss:   2.52, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(4.2021017, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   10/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.06, prob_loss:   2.05, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(3.8746715, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   11/125, lr:0.000003, giou_loss:   1.75, conf_loss:   0.02, prob_loss:   2.11, total_loss:   3.87\n",
            " Total Loss  tf.Tensor(4.694025, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   12/125, lr:0.000003, giou_loss:   2.25, conf_loss:   0.01, prob_loss:   2.43, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(5.054537, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   13/125, lr:0.000003, giou_loss:   2.24, conf_loss:   0.47, prob_loss:   2.34, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(3.5460963, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   14/125, lr:0.000003, giou_loss:   1.65, conf_loss:   0.02, prob_loss:   1.87, total_loss:   3.55\n",
            " Total Loss  tf.Tensor(4.8141146, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   15/125, lr:0.000003, giou_loss:   2.32, conf_loss:   0.32, prob_loss:   2.18, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(4.013611, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   16/125, lr:0.000003, giou_loss:   1.83, conf_loss:   0.17, prob_loss:   2.01, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(3.8266258, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   17/125, lr:0.000003, giou_loss:   1.93, conf_loss:   0.03, prob_loss:   1.87, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(4.7888536, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   18/125, lr:0.000003, giou_loss:   2.22, conf_loss:   0.10, prob_loss:   2.47, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(3.5416462, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   19/125, lr:0.000003, giou_loss:   1.53, conf_loss:   0.20, prob_loss:   1.81, total_loss:   3.54\n",
            " Total Loss  tf.Tensor(6.371278, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   20/125, lr:0.000003, giou_loss:   3.18, conf_loss:   0.06, prob_loss:   3.14, total_loss:   6.37\n",
            " Total Loss  tf.Tensor(5.0068665, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   21/125, lr:0.000003, giou_loss:   2.64, conf_loss:   0.04, prob_loss:   2.32, total_loss:   5.01\n",
            " Total Loss  tf.Tensor(4.330306, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   22/125, lr:0.000003, giou_loss:   1.92, conf_loss:   0.01, prob_loss:   2.39, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(4.979678, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   23/125, lr:0.000003, giou_loss:   2.23, conf_loss:   0.06, prob_loss:   2.69, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(4.723817, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   24/125, lr:0.000003, giou_loss:   2.20, conf_loss:   0.14, prob_loss:   2.38, total_loss:   4.72\n",
            " Total Loss  tf.Tensor(6.621377, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   25/125, lr:0.000003, giou_loss:   3.31, conf_loss:   0.16, prob_loss:   3.15, total_loss:   6.62\n",
            " Total Loss  tf.Tensor(5.011922, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   26/125, lr:0.000003, giou_loss:   2.40, conf_loss:   0.08, prob_loss:   2.53, total_loss:   5.01\n",
            " Total Loss  tf.Tensor(5.2152843, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   27/125, lr:0.000003, giou_loss:   2.35, conf_loss:   0.04, prob_loss:   2.83, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(3.3502383, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   28/125, lr:0.000003, giou_loss:   1.63, conf_loss:   0.07, prob_loss:   1.65, total_loss:   3.35\n",
            " Total Loss  tf.Tensor(4.4542294, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   29/125, lr:0.000003, giou_loss:   2.15, conf_loss:   0.08, prob_loss:   2.23, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.974939, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   30/125, lr:0.000003, giou_loss:   2.31, conf_loss:   0.04, prob_loss:   2.63, total_loss:   4.97\n",
            " Total Loss  tf.Tensor(5.3505077, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   31/125, lr:0.000003, giou_loss:   2.66, conf_loss:   0.09, prob_loss:   2.59, total_loss:   5.35\n",
            " Total Loss  tf.Tensor(4.939693, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   32/125, lr:0.000003, giou_loss:   2.41, conf_loss:   0.12, prob_loss:   2.41, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(3.7002687, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   33/125, lr:0.000003, giou_loss:   1.70, conf_loss:   0.07, prob_loss:   1.93, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(5.216894, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   34/125, lr:0.000003, giou_loss:   2.55, conf_loss:   0.08, prob_loss:   2.59, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(5.31866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   35/125, lr:0.000003, giou_loss:   2.46, conf_loss:   0.06, prob_loss:   2.80, total_loss:   5.32\n",
            " Total Loss  tf.Tensor(4.8406687, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   36/125, lr:0.000003, giou_loss:   2.27, conf_loss:   0.06, prob_loss:   2.50, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(5.9713445, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   37/125, lr:0.000003, giou_loss:   3.08, conf_loss:   0.23, prob_loss:   2.66, total_loss:   5.97\n",
            " Total Loss  tf.Tensor(4.414268, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   38/125, lr:0.000003, giou_loss:   2.16, conf_loss:   0.05, prob_loss:   2.20, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(4.405613, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   39/125, lr:0.000003, giou_loss:   2.24, conf_loss:   0.05, prob_loss:   2.11, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(4.6524887, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   40/125, lr:0.000003, giou_loss:   2.02, conf_loss:   0.09, prob_loss:   2.54, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(4.05417, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   41/125, lr:0.000003, giou_loss:   1.80, conf_loss:   0.02, prob_loss:   2.23, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(3.3209538, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   42/125, lr:0.000003, giou_loss:   1.62, conf_loss:   0.03, prob_loss:   1.68, total_loss:   3.32\n",
            " Total Loss  tf.Tensor(5.8359823, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   43/125, lr:0.000003, giou_loss:   3.00, conf_loss:   0.18, prob_loss:   2.66, total_loss:   5.84\n",
            " Total Loss  tf.Tensor(5.518616, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   44/125, lr:0.000003, giou_loss:   2.74, conf_loss:   0.09, prob_loss:   2.69, total_loss:   5.52\n",
            " Total Loss  tf.Tensor(4.8983727, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   45/125, lr:0.000003, giou_loss:   2.25, conf_loss:   0.15, prob_loss:   2.49, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(5.62976, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   46/125, lr:0.000003, giou_loss:   2.64, conf_loss:   0.10, prob_loss:   2.90, total_loss:   5.63\n",
            " Total Loss  tf.Tensor(5.132643, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   47/125, lr:0.000003, giou_loss:   2.27, conf_loss:   0.14, prob_loss:   2.72, total_loss:   5.13\n",
            " Total Loss  tf.Tensor(3.9852698, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   48/125, lr:0.000003, giou_loss:   1.95, conf_loss:   0.15, prob_loss:   1.88, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(6.079504, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   49/125, lr:0.000003, giou_loss:   2.88, conf_loss:   0.06, prob_loss:   3.14, total_loss:   6.08\n",
            " Total Loss  tf.Tensor(5.117815, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   50/125, lr:0.000003, giou_loss:   2.32, conf_loss:   0.05, prob_loss:   2.74, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(4.3429604, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   51/125, lr:0.000003, giou_loss:   2.03, conf_loss:   0.05, prob_loss:   2.26, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(5.1736593, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   52/125, lr:0.000003, giou_loss:   2.67, conf_loss:   0.06, prob_loss:   2.45, total_loss:   5.17\n",
            " Total Loss  tf.Tensor(4.83603, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   53/125, lr:0.000003, giou_loss:   2.28, conf_loss:   0.05, prob_loss:   2.51, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.0133076, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   54/125, lr:0.000003, giou_loss:   1.78, conf_loss:   0.19, prob_loss:   2.04, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(4.361017, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   55/125, lr:0.000003, giou_loss:   2.20, conf_loss:   0.05, prob_loss:   2.12, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.585694, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   56/125, lr:0.000003, giou_loss:   2.29, conf_loss:   0.21, prob_loss:   2.08, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(5.0016646, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   57/125, lr:0.000003, giou_loss:   2.21, conf_loss:   0.04, prob_loss:   2.75, total_loss:   5.00\n",
            " Total Loss  tf.Tensor(3.6548626, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   58/125, lr:0.000003, giou_loss:   1.54, conf_loss:   0.11, prob_loss:   2.01, total_loss:   3.65\n",
            " Total Loss  tf.Tensor(5.21821, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   59/125, lr:0.000003, giou_loss:   2.20, conf_loss:   0.11, prob_loss:   2.91, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(4.5608215, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   60/125, lr:0.000003, giou_loss:   2.00, conf_loss:   0.22, prob_loss:   2.33, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.238453, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   61/125, lr:0.000003, giou_loss:   1.98, conf_loss:   0.02, prob_loss:   2.23, total_loss:   4.24\n",
            " Total Loss  tf.Tensor(4.729518, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   62/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.03, prob_loss:   2.61, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(4.58644, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   63/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.05, prob_loss:   2.44, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(3.8916173, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   64/125, lr:0.000003, giou_loss:   1.53, conf_loss:   0.12, prob_loss:   2.25, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(4.4293575, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   65/125, lr:0.000003, giou_loss:   2.00, conf_loss:   0.19, prob_loss:   2.24, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(3.8280866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   66/125, lr:0.000003, giou_loss:   1.92, conf_loss:   0.12, prob_loss:   1.79, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(4.2733736, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   67/125, lr:0.000003, giou_loss:   1.90, conf_loss:   0.04, prob_loss:   2.33, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(4.3383694, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   68/125, lr:0.000003, giou_loss:   1.81, conf_loss:   0.22, prob_loss:   2.30, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(4.355323, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   69/125, lr:0.000003, giou_loss:   2.02, conf_loss:   0.10, prob_loss:   2.24, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.034813, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   70/125, lr:0.000003, giou_loss:   1.92, conf_loss:   0.02, prob_loss:   2.09, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(4.1509695, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   71/125, lr:0.000003, giou_loss:   2.06, conf_loss:   0.14, prob_loss:   1.96, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(3.914744, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   72/125, lr:0.000003, giou_loss:   1.88, conf_loss:   0.03, prob_loss:   2.00, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.543853, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   73/125, lr:0.000003, giou_loss:   2.21, conf_loss:   0.04, prob_loss:   2.30, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(4.596326, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   74/125, lr:0.000003, giou_loss:   2.13, conf_loss:   0.12, prob_loss:   2.34, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(5.1642513, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   75/125, lr:0.000003, giou_loss:   2.39, conf_loss:   0.03, prob_loss:   2.75, total_loss:   5.16\n",
            " Total Loss  tf.Tensor(3.813074, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   76/125, lr:0.000003, giou_loss:   1.82, conf_loss:   0.08, prob_loss:   1.92, total_loss:   3.81\n",
            " Total Loss  tf.Tensor(4.7562866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   77/125, lr:0.000003, giou_loss:   2.23, conf_loss:   0.16, prob_loss:   2.37, total_loss:   4.76\n",
            " Total Loss  tf.Tensor(4.111385, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   78/125, lr:0.000003, giou_loss:   1.83, conf_loss:   0.05, prob_loss:   2.23, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(5.548914, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   79/125, lr:0.000003, giou_loss:   2.48, conf_loss:   0.06, prob_loss:   3.01, total_loss:   5.55\n",
            " Total Loss  tf.Tensor(4.2497025, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   80/125, lr:0.000003, giou_loss:   2.05, conf_loss:   0.06, prob_loss:   2.14, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(5.408872, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   81/125, lr:0.000003, giou_loss:   2.71, conf_loss:   0.06, prob_loss:   2.64, total_loss:   5.41\n",
            " Total Loss  tf.Tensor(3.9391541, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   82/125, lr:0.000003, giou_loss:   1.94, conf_loss:   0.03, prob_loss:   1.97, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(4.7201595, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   83/125, lr:0.000003, giou_loss:   2.37, conf_loss:   0.02, prob_loss:   2.33, total_loss:   4.72\n",
            " Total Loss  tf.Tensor(6.048745, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   84/125, lr:0.000003, giou_loss:   2.84, conf_loss:   0.18, prob_loss:   3.03, total_loss:   6.05\n",
            " Total Loss  tf.Tensor(5.1397953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   85/125, lr:0.000003, giou_loss:   2.28, conf_loss:   0.14, prob_loss:   2.73, total_loss:   5.14\n",
            " Total Loss  tf.Tensor(5.6207447, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   86/125, lr:0.000003, giou_loss:   2.73, conf_loss:   0.14, prob_loss:   2.75, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(3.8820953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   87/125, lr:0.000003, giou_loss:   2.03, conf_loss:   0.10, prob_loss:   1.76, total_loss:   3.88\n",
            " Total Loss  tf.Tensor(4.7514067, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   88/125, lr:0.000003, giou_loss:   2.04, conf_loss:   0.38, prob_loss:   2.33, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(4.303253, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   89/125, lr:0.000003, giou_loss:   2.07, conf_loss:   0.09, prob_loss:   2.14, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(2.7384417, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   90/125, lr:0.000003, giou_loss:   1.30, conf_loss:   0.10, prob_loss:   1.34, total_loss:   2.74\n",
            " Total Loss  tf.Tensor(4.063635, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   91/125, lr:0.000003, giou_loss:   1.84, conf_loss:   0.02, prob_loss:   2.20, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(5.972963, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   92/125, lr:0.000003, giou_loss:   2.97, conf_loss:   0.07, prob_loss:   2.94, total_loss:   5.97\n",
            " Total Loss  tf.Tensor(3.415104, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   93/125, lr:0.000003, giou_loss:   1.60, conf_loss:   0.01, prob_loss:   1.80, total_loss:   3.42\n",
            " Total Loss  tf.Tensor(3.514227, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   94/125, lr:0.000003, giou_loss:   1.63, conf_loss:   0.02, prob_loss:   1.86, total_loss:   3.51\n",
            " Total Loss  tf.Tensor(4.7852345, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   95/125, lr:0.000003, giou_loss:   2.19, conf_loss:   0.08, prob_loss:   2.52, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.436749, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   96/125, lr:0.000003, giou_loss:   2.16, conf_loss:   0.05, prob_loss:   2.23, total_loss:   4.44\n",
            " Total Loss  tf.Tensor(4.2106853, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   97/125, lr:0.000003, giou_loss:   1.79, conf_loss:   0.03, prob_loss:   2.39, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(3.6118317, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   98/125, lr:0.000003, giou_loss:   1.65, conf_loss:   0.03, prob_loss:   1.93, total_loss:   3.61\n",
            " Total Loss  tf.Tensor(3.3623972, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:   99/125, lr:0.000003, giou_loss:   1.51, conf_loss:   0.02, prob_loss:   1.83, total_loss:   3.36\n",
            " Total Loss  tf.Tensor(5.3126287, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  100/125, lr:0.000003, giou_loss:   2.52, conf_loss:   0.14, prob_loss:   2.65, total_loss:   5.31\n",
            " Total Loss  tf.Tensor(4.670826, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  101/125, lr:0.000003, giou_loss:   2.27, conf_loss:   0.05, prob_loss:   2.35, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.585602, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  102/125, lr:0.000003, giou_loss:   2.83, conf_loss:   0.10, prob_loss:   2.65, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(4.531763, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  103/125, lr:0.000003, giou_loss:   2.20, conf_loss:   0.16, prob_loss:   2.17, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(3.972803, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  104/125, lr:0.000003, giou_loss:   1.77, conf_loss:   0.05, prob_loss:   2.15, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(3.940641, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  105/125, lr:0.000003, giou_loss:   1.68, conf_loss:   0.08, prob_loss:   2.18, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(6.06627, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  106/125, lr:0.000003, giou_loss:   2.44, conf_loss:   0.50, prob_loss:   3.12, total_loss:   6.07\n",
            " Total Loss  tf.Tensor(5.2725363, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  107/125, lr:0.000003, giou_loss:   2.51, conf_loss:   0.25, prob_loss:   2.51, total_loss:   5.27\n",
            " Total Loss  tf.Tensor(4.6439304, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  108/125, lr:0.000003, giou_loss:   2.17, conf_loss:   0.10, prob_loss:   2.38, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(3.8593588, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  109/125, lr:0.000003, giou_loss:   1.72, conf_loss:   0.11, prob_loss:   2.03, total_loss:   3.86\n",
            " Total Loss  tf.Tensor(4.4125605, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  110/125, lr:0.000003, giou_loss:   2.14, conf_loss:   0.02, prob_loss:   2.25, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(4.5890017, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  111/125, lr:0.000003, giou_loss:   2.28, conf_loss:   0.04, prob_loss:   2.27, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(3.677496, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  112/125, lr:0.000003, giou_loss:   1.79, conf_loss:   0.01, prob_loss:   1.88, total_loss:   3.68\n",
            " Total Loss  tf.Tensor(4.60464, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  113/125, lr:0.000003, giou_loss:   2.12, conf_loss:   0.04, prob_loss:   2.44, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(4.723263, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  114/125, lr:0.000003, giou_loss:   2.14, conf_loss:   0.10, prob_loss:   2.48, total_loss:   4.72\n",
            " Total Loss  tf.Tensor(4.823552, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  115/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.41, prob_loss:   2.33, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(3.8316088, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  116/125, lr:0.000003, giou_loss:   1.86, conf_loss:   0.04, prob_loss:   1.93, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(3.413834, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  117/125, lr:0.000003, giou_loss:   1.73, conf_loss:   0.11, prob_loss:   1.58, total_loss:   3.41\n",
            " Total Loss  tf.Tensor(5.0222855, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  118/125, lr:0.000003, giou_loss:   2.26, conf_loss:   0.11, prob_loss:   2.65, total_loss:   5.02\n",
            " Total Loss  tf.Tensor(3.7706037, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  119/125, lr:0.000003, giou_loss:   1.67, conf_loss:   0.10, prob_loss:   2.00, total_loss:   3.77\n",
            " Total Loss  tf.Tensor(5.6158752, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  120/125, lr:0.000003, giou_loss:   1.94, conf_loss:   0.11, prob_loss:   3.57, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(4.031584, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  121/125, lr:0.000003, giou_loss:   1.82, conf_loss:   0.04, prob_loss:   2.17, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(4.232179, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  122/125, lr:0.000003, giou_loss:   1.93, conf_loss:   0.13, prob_loss:   2.17, total_loss:   4.23\n",
            " Total Loss  tf.Tensor(3.7846007, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  123/125, lr:0.000003, giou_loss:   1.67, conf_loss:   0.04, prob_loss:   2.07, total_loss:   3.78\n",
            " Total Loss  tf.Tensor(4.862178, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:  124/125, lr:0.000003, giou_loss:   2.25, conf_loss:   0.21, prob_loss:   2.40, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(3.86047, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    0/125, lr:0.000003, giou_loss:   1.75, conf_loss:   0.02, prob_loss:   2.09, total_loss:   3.86\n",
            " Total Loss  tf.Tensor(4.832048, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:90 step:    1/125, lr:0.000003, giou_loss:   2.21, conf_loss:   0.07, prob_loss:   2.55, total_loss:   4.83\n",
            "\n",
            "\n",
            "giou_val_loss:   2.25, conf_val_loss:   0.64, prob_val_loss:   5.25, total_val_loss:   8.14\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(3.710949, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    2/125, lr:0.000003, giou_loss:   1.59, conf_loss:   0.25, prob_loss:   1.87, total_loss:   3.71\n",
            " Total Loss  tf.Tensor(5.176063, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    3/125, lr:0.000003, giou_loss:   2.45, conf_loss:   0.14, prob_loss:   2.59, total_loss:   5.18\n",
            " Total Loss  tf.Tensor(4.455495, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    4/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.06, prob_loss:   2.31, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(5.0605793, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    5/125, lr:0.000003, giou_loss:   2.45, conf_loss:   0.07, prob_loss:   2.54, total_loss:   5.06\n",
            " Total Loss  tf.Tensor(5.8195305, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    6/125, lr:0.000003, giou_loss:   2.81, conf_loss:   0.14, prob_loss:   2.87, total_loss:   5.82\n",
            " Total Loss  tf.Tensor(4.077127, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    7/125, lr:0.000003, giou_loss:   2.07, conf_loss:   0.09, prob_loss:   1.91, total_loss:   4.08\n",
            " Total Loss  tf.Tensor(4.851183, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    8/125, lr:0.000003, giou_loss:   2.32, conf_loss:   0.03, prob_loss:   2.50, total_loss:   4.85\n",
            " Total Loss  tf.Tensor(4.810982, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    9/125, lr:0.000003, giou_loss:   2.12, conf_loss:   0.11, prob_loss:   2.58, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(5.3923607, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   10/125, lr:0.000003, giou_loss:   2.54, conf_loss:   0.29, prob_loss:   2.56, total_loss:   5.39\n",
            " Total Loss  tf.Tensor(3.8341584, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   11/125, lr:0.000003, giou_loss:   1.85, conf_loss:   0.06, prob_loss:   1.92, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(4.632124, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   12/125, lr:0.000003, giou_loss:   2.07, conf_loss:   0.23, prob_loss:   2.33, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(4.9513674, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   13/125, lr:0.000003, giou_loss:   2.40, conf_loss:   0.14, prob_loss:   2.41, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(5.158146, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   14/125, lr:0.000003, giou_loss:   2.37, conf_loss:   0.04, prob_loss:   2.75, total_loss:   5.16\n",
            " Total Loss  tf.Tensor(5.1937695, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   15/125, lr:0.000003, giou_loss:   2.49, conf_loss:   0.13, prob_loss:   2.57, total_loss:   5.19\n",
            " Total Loss  tf.Tensor(4.520008, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   16/125, lr:0.000003, giou_loss:   2.04, conf_loss:   0.04, prob_loss:   2.44, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(4.8777866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   17/125, lr:0.000003, giou_loss:   2.46, conf_loss:   0.11, prob_loss:   2.30, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(4.392508, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   18/125, lr:0.000003, giou_loss:   2.00, conf_loss:   0.09, prob_loss:   2.30, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(5.827559, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   19/125, lr:0.000003, giou_loss:   2.85, conf_loss:   0.06, prob_loss:   2.92, total_loss:   5.83\n",
            " Total Loss  tf.Tensor(4.5846634, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   20/125, lr:0.000003, giou_loss:   2.19, conf_loss:   0.05, prob_loss:   2.34, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.964643, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   21/125, lr:0.000003, giou_loss:   1.75, conf_loss:   0.11, prob_loss:   2.10, total_loss:   3.96\n",
            " Total Loss  tf.Tensor(5.15197, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   22/125, lr:0.000003, giou_loss:   2.26, conf_loss:   0.12, prob_loss:   2.77, total_loss:   5.15\n",
            " Total Loss  tf.Tensor(4.20494, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   23/125, lr:0.000003, giou_loss:   1.92, conf_loss:   0.08, prob_loss:   2.20, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(5.072048, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   24/125, lr:0.000003, giou_loss:   2.30, conf_loss:   0.06, prob_loss:   2.72, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(3.6914072, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   25/125, lr:0.000003, giou_loss:   1.64, conf_loss:   0.02, prob_loss:   2.03, total_loss:   3.69\n",
            " Total Loss  tf.Tensor(5.084655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   26/125, lr:0.000003, giou_loss:   2.29, conf_loss:   0.53, prob_loss:   2.26, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(5.073789, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   27/125, lr:0.000003, giou_loss:   2.49, conf_loss:   0.03, prob_loss:   2.56, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(4.5989237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   28/125, lr:0.000003, giou_loss:   2.15, conf_loss:   0.02, prob_loss:   2.43, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(3.9064574, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   29/125, lr:0.000003, giou_loss:   1.87, conf_loss:   0.09, prob_loss:   1.95, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(5.252657, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   30/125, lr:0.000003, giou_loss:   2.45, conf_loss:   0.08, prob_loss:   2.72, total_loss:   5.25\n",
            " Total Loss  tf.Tensor(3.7454758, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   31/125, lr:0.000003, giou_loss:   1.48, conf_loss:   0.01, prob_loss:   2.25, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(4.4363737, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   32/125, lr:0.000003, giou_loss:   2.03, conf_loss:   0.05, prob_loss:   2.35, total_loss:   4.44\n",
            " Total Loss  tf.Tensor(4.64744, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   33/125, lr:0.000003, giou_loss:   2.05, conf_loss:   0.24, prob_loss:   2.35, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(4.3892727, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   34/125, lr:0.000003, giou_loss:   1.94, conf_loss:   0.10, prob_loss:   2.35, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(3.8220162, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   35/125, lr:0.000003, giou_loss:   1.95, conf_loss:   0.06, prob_loss:   1.82, total_loss:   3.82\n",
            " Total Loss  tf.Tensor(4.500891, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   36/125, lr:0.000003, giou_loss:   1.94, conf_loss:   0.10, prob_loss:   2.46, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(3.692998, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   37/125, lr:0.000003, giou_loss:   1.69, conf_loss:   0.09, prob_loss:   1.91, total_loss:   3.69\n",
            " Total Loss  tf.Tensor(4.20152, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   38/125, lr:0.000003, giou_loss:   1.84, conf_loss:   0.04, prob_loss:   2.32, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(4.182617, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   39/125, lr:0.000003, giou_loss:   1.97, conf_loss:   0.04, prob_loss:   2.16, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(4.6534133, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   40/125, lr:0.000003, giou_loss:   2.01, conf_loss:   0.01, prob_loss:   2.63, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(3.201911, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   41/125, lr:0.000003, giou_loss:   1.42, conf_loss:   0.03, prob_loss:   1.76, total_loss:   3.20\n",
            " Total Loss  tf.Tensor(3.377161, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   42/125, lr:0.000003, giou_loss:   1.34, conf_loss:   0.44, prob_loss:   1.60, total_loss:   3.38\n",
            " Total Loss  tf.Tensor(4.7148933, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   43/125, lr:0.000003, giou_loss:   2.16, conf_loss:   0.06, prob_loss:   2.49, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(4.655739, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   44/125, lr:0.000003, giou_loss:   2.10, conf_loss:   0.13, prob_loss:   2.43, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(3.429294, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   45/125, lr:0.000003, giou_loss:   1.61, conf_loss:   0.02, prob_loss:   1.80, total_loss:   3.43\n",
            " Total Loss  tf.Tensor(4.3509455, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   46/125, lr:0.000003, giou_loss:   1.99, conf_loss:   0.23, prob_loss:   2.13, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(4.1654077, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   47/125, lr:0.000003, giou_loss:   2.01, conf_loss:   0.03, prob_loss:   2.12, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(5.5463333, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   48/125, lr:0.000003, giou_loss:   2.56, conf_loss:   0.05, prob_loss:   2.93, total_loss:   5.55\n",
            " Total Loss  tf.Tensor(4.1983395, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   49/125, lr:0.000003, giou_loss:   1.87, conf_loss:   0.20, prob_loss:   2.13, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(4.682464, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   50/125, lr:0.000003, giou_loss:   2.20, conf_loss:   0.04, prob_loss:   2.44, total_loss:   4.68\n",
            " Total Loss  tf.Tensor(4.050805, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   51/125, lr:0.000003, giou_loss:   1.85, conf_loss:   0.08, prob_loss:   2.12, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(4.538186, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   52/125, lr:0.000003, giou_loss:   2.29, conf_loss:   0.03, prob_loss:   2.21, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(4.3251257, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   53/125, lr:0.000003, giou_loss:   2.04, conf_loss:   0.13, prob_loss:   2.15, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(4.367447, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   54/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.02, prob_loss:   2.26, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(5.5359335, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   55/125, lr:0.000003, giou_loss:   2.56, conf_loss:   0.39, prob_loss:   2.58, total_loss:   5.54\n",
            " Total Loss  tf.Tensor(4.108033, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   56/125, lr:0.000003, giou_loss:   1.97, conf_loss:   0.03, prob_loss:   2.10, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(4.7935123, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   57/125, lr:0.000003, giou_loss:   2.31, conf_loss:   0.03, prob_loss:   2.45, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.057673, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   58/125, lr:0.000003, giou_loss:   1.85, conf_loss:   0.07, prob_loss:   2.15, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(5.3132563, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   59/125, lr:0.000003, giou_loss:   2.65, conf_loss:   0.05, prob_loss:   2.62, total_loss:   5.31\n",
            " Total Loss  tf.Tensor(4.194976, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   60/125, lr:0.000003, giou_loss:   1.86, conf_loss:   0.10, prob_loss:   2.24, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(4.9843636, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   61/125, lr:0.000003, giou_loss:   2.54, conf_loss:   0.08, prob_loss:   2.36, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(4.7854233, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   62/125, lr:0.000003, giou_loss:   2.35, conf_loss:   0.03, prob_loss:   2.41, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.50002, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   63/125, lr:0.000003, giou_loss:   2.07, conf_loss:   0.11, prob_loss:   2.32, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.355032, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   64/125, lr:0.000003, giou_loss:   2.08, conf_loss:   0.06, prob_loss:   2.22, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(3.9233081, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   65/125, lr:0.000003, giou_loss:   1.88, conf_loss:   0.09, prob_loss:   1.95, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(5.686795, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   66/125, lr:0.000003, giou_loss:   2.77, conf_loss:   0.07, prob_loss:   2.84, total_loss:   5.69\n",
            " Total Loss  tf.Tensor(5.588663, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   67/125, lr:0.000003, giou_loss:   2.59, conf_loss:   0.22, prob_loss:   2.78, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(4.8388124, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   68/125, lr:0.000003, giou_loss:   2.36, conf_loss:   0.09, prob_loss:   2.38, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.979945, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   69/125, lr:0.000003, giou_loss:   2.21, conf_loss:   0.06, prob_loss:   2.71, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(4.572653, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   70/125, lr:0.000003, giou_loss:   2.18, conf_loss:   0.05, prob_loss:   2.35, total_loss:   4.57\n",
            " Total Loss  tf.Tensor(4.038079, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   71/125, lr:0.000003, giou_loss:   1.91, conf_loss:   0.04, prob_loss:   2.08, total_loss:   4.04\n",
            " Total Loss  tf.Tensor(4.766462, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   72/125, lr:0.000003, giou_loss:   2.09, conf_loss:   0.04, prob_loss:   2.63, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(3.7637954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   73/125, lr:0.000003, giou_loss:   1.69, conf_loss:   0.11, prob_loss:   1.96, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(4.3623466, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   74/125, lr:0.000003, giou_loss:   2.00, conf_loss:   0.12, prob_loss:   2.24, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(5.02174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   75/125, lr:0.000003, giou_loss:   2.36, conf_loss:   0.10, prob_loss:   2.56, total_loss:   5.02\n",
            " Total Loss  tf.Tensor(4.8084173, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   76/125, lr:0.000003, giou_loss:   2.29, conf_loss:   0.07, prob_loss:   2.45, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(4.4677324, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   77/125, lr:0.000003, giou_loss:   2.01, conf_loss:   0.06, prob_loss:   2.40, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(5.5245285, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   78/125, lr:0.000003, giou_loss:   2.67, conf_loss:   0.14, prob_loss:   2.72, total_loss:   5.52\n",
            " Total Loss  tf.Tensor(4.2990303, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   79/125, lr:0.000003, giou_loss:   2.05, conf_loss:   0.05, prob_loss:   2.20, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(3.9113307, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   80/125, lr:0.000003, giou_loss:   1.90, conf_loss:   0.10, prob_loss:   1.92, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.0508175, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   81/125, lr:0.000003, giou_loss:   1.71, conf_loss:   0.10, prob_loss:   2.24, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(3.7626953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   82/125, lr:0.000003, giou_loss:   1.78, conf_loss:   0.02, prob_loss:   1.96, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(4.177223, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   83/125, lr:0.000003, giou_loss:   1.81, conf_loss:   0.03, prob_loss:   2.34, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(3.494358, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   84/125, lr:0.000003, giou_loss:   1.72, conf_loss:   0.09, prob_loss:   1.69, total_loss:   3.49\n",
            " Total Loss  tf.Tensor(5.5868835, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   85/125, lr:0.000003, giou_loss:   2.41, conf_loss:   0.13, prob_loss:   3.05, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(4.7862153, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   86/125, lr:0.000003, giou_loss:   2.30, conf_loss:   0.06, prob_loss:   2.43, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.3397512, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   87/125, lr:0.000003, giou_loss:   2.16, conf_loss:   0.07, prob_loss:   2.11, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(4.185685, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   88/125, lr:0.000003, giou_loss:   1.88, conf_loss:   0.02, prob_loss:   2.28, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(5.5853944, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   89/125, lr:0.000003, giou_loss:   2.69, conf_loss:   0.23, prob_loss:   2.66, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(4.187337, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   90/125, lr:0.000003, giou_loss:   1.99, conf_loss:   0.10, prob_loss:   2.11, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(3.9904342, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   91/125, lr:0.000003, giou_loss:   1.83, conf_loss:   0.06, prob_loss:   2.11, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(2.7781272, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   92/125, lr:0.000003, giou_loss:   1.36, conf_loss:   0.02, prob_loss:   1.40, total_loss:   2.78\n",
            " Total Loss  tf.Tensor(4.1661587, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   93/125, lr:0.000003, giou_loss:   2.02, conf_loss:   0.02, prob_loss:   2.12, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.356003, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   94/125, lr:0.000003, giou_loss:   2.06, conf_loss:   0.01, prob_loss:   2.29, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.9492464, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   95/125, lr:0.000003, giou_loss:   2.49, conf_loss:   0.15, prob_loss:   2.31, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(4.693117, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   96/125, lr:0.000003, giou_loss:   2.33, conf_loss:   0.02, prob_loss:   2.35, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(5.1082134, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   97/125, lr:0.000003, giou_loss:   2.26, conf_loss:   0.31, prob_loss:   2.54, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(5.1165285, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   98/125, lr:0.000003, giou_loss:   2.52, conf_loss:   0.11, prob_loss:   2.49, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(4.3182335, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:   99/125, lr:0.000003, giou_loss:   2.07, conf_loss:   0.05, prob_loss:   2.20, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(5.334592, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  100/125, lr:0.000003, giou_loss:   2.63, conf_loss:   0.12, prob_loss:   2.58, total_loss:   5.33\n",
            " Total Loss  tf.Tensor(4.282645, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  101/125, lr:0.000003, giou_loss:   1.88, conf_loss:   0.12, prob_loss:   2.28, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(4.60445, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  102/125, lr:0.000003, giou_loss:   2.13, conf_loss:   0.15, prob_loss:   2.32, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(4.2057967, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  103/125, lr:0.000003, giou_loss:   1.91, conf_loss:   0.04, prob_loss:   2.25, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(3.1563396, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  104/125, lr:0.000003, giou_loss:   1.49, conf_loss:   0.25, prob_loss:   1.41, total_loss:   3.16\n",
            " Total Loss  tf.Tensor(5.617279, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  105/125, lr:0.000003, giou_loss:   2.52, conf_loss:   0.06, prob_loss:   3.04, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(4.5551057, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  106/125, lr:0.000003, giou_loss:   2.00, conf_loss:   0.03, prob_loss:   2.52, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(5.6288233, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  107/125, lr:0.000003, giou_loss:   2.64, conf_loss:   0.17, prob_loss:   2.81, total_loss:   5.63\n",
            " Total Loss  tf.Tensor(5.4599905, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  108/125, lr:0.000003, giou_loss:   2.56, conf_loss:   0.16, prob_loss:   2.74, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(5.5217156, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  109/125, lr:0.000003, giou_loss:   2.44, conf_loss:   0.20, prob_loss:   2.89, total_loss:   5.52\n",
            " Total Loss  tf.Tensor(2.9906325, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  110/125, lr:0.000003, giou_loss:   1.34, conf_loss:   0.03, prob_loss:   1.61, total_loss:   2.99\n",
            " Total Loss  tf.Tensor(4.4698834, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  111/125, lr:0.000003, giou_loss:   1.91, conf_loss:   0.10, prob_loss:   2.47, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(4.8647046, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  112/125, lr:0.000003, giou_loss:   2.38, conf_loss:   0.06, prob_loss:   2.43, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(3.6999648, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  113/125, lr:0.000003, giou_loss:   1.73, conf_loss:   0.06, prob_loss:   1.90, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(5.0426893, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  114/125, lr:0.000003, giou_loss:   2.37, conf_loss:   0.10, prob_loss:   2.58, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(4.3238106, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  115/125, lr:0.000003, giou_loss:   2.04, conf_loss:   0.17, prob_loss:   2.12, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(3.5074, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  116/125, lr:0.000003, giou_loss:   1.49, conf_loss:   0.02, prob_loss:   2.00, total_loss:   3.51\n",
            " Total Loss  tf.Tensor(4.278352, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  117/125, lr:0.000003, giou_loss:   1.98, conf_loss:   0.02, prob_loss:   2.28, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(4.394432, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  118/125, lr:0.000003, giou_loss:   1.99, conf_loss:   0.08, prob_loss:   2.33, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(4.791397, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  119/125, lr:0.000003, giou_loss:   2.16, conf_loss:   0.07, prob_loss:   2.56, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.4827685, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  120/125, lr:0.000003, giou_loss:   1.97, conf_loss:   0.04, prob_loss:   2.48, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(3.7960744, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  121/125, lr:0.000003, giou_loss:   1.78, conf_loss:   0.03, prob_loss:   1.98, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(4.330699, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  122/125, lr:0.000003, giou_loss:   2.11, conf_loss:   0.03, prob_loss:   2.19, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(5.102583, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  123/125, lr:0.000003, giou_loss:   2.22, conf_loss:   0.18, prob_loss:   2.70, total_loss:   5.10\n",
            " Total Loss  tf.Tensor(3.6569552, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:  124/125, lr:0.000003, giou_loss:   1.73, conf_loss:   0.08, prob_loss:   1.85, total_loss:   3.66\n",
            " Total Loss  tf.Tensor(4.2472835, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    0/125, lr:0.000003, giou_loss:   2.12, conf_loss:   0.06, prob_loss:   2.07, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(4.5985317, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:91 step:    1/125, lr:0.000003, giou_loss:   2.11, conf_loss:   0.06, prob_loss:   2.43, total_loss:   4.60\n",
            "\n",
            "\n",
            "giou_val_loss:   2.20, conf_val_loss:   0.63, prob_val_loss:   5.19, total_val_loss:   8.02\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.733266, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    2/125, lr:0.000003, giou_loss:   2.19, conf_loss:   0.21, prob_loss:   2.33, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(4.7103357, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    3/125, lr:0.000003, giou_loss:   2.00, conf_loss:   0.08, prob_loss:   2.63, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(5.702469, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    4/125, lr:0.000003, giou_loss:   2.62, conf_loss:   0.62, prob_loss:   2.46, total_loss:   5.70\n",
            " Total Loss  tf.Tensor(4.470729, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    5/125, lr:0.000003, giou_loss:   2.14, conf_loss:   0.17, prob_loss:   2.16, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(3.4565606, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    6/125, lr:0.000003, giou_loss:   1.56, conf_loss:   0.01, prob_loss:   1.89, total_loss:   3.46\n",
            " Total Loss  tf.Tensor(5.2508917, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    7/125, lr:0.000003, giou_loss:   2.59, conf_loss:   0.08, prob_loss:   2.58, total_loss:   5.25\n",
            " Total Loss  tf.Tensor(4.196471, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    8/125, lr:0.000003, giou_loss:   1.85, conf_loss:   0.05, prob_loss:   2.30, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(4.788534, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    9/125, lr:0.000003, giou_loss:   2.17, conf_loss:   0.07, prob_loss:   2.55, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.7066345, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   10/125, lr:0.000003, giou_loss:   2.39, conf_loss:   0.11, prob_loss:   2.20, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(5.014265, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   11/125, lr:0.000003, giou_loss:   2.59, conf_loss:   0.07, prob_loss:   2.35, total_loss:   5.01\n",
            " Total Loss  tf.Tensor(3.5585184, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   12/125, lr:0.000003, giou_loss:   1.60, conf_loss:   0.02, prob_loss:   1.94, total_loss:   3.56\n",
            " Total Loss  tf.Tensor(4.683198, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   13/125, lr:0.000003, giou_loss:   2.11, conf_loss:   0.08, prob_loss:   2.49, total_loss:   4.68\n",
            " Total Loss  tf.Tensor(4.8863497, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   14/125, lr:0.000003, giou_loss:   2.13, conf_loss:   0.13, prob_loss:   2.63, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(4.4744177, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   15/125, lr:0.000003, giou_loss:   2.14, conf_loss:   0.03, prob_loss:   2.30, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(4.0983515, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   16/125, lr:0.000003, giou_loss:   1.87, conf_loss:   0.05, prob_loss:   2.18, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(4.6521997, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   17/125, lr:0.000003, giou_loss:   2.21, conf_loss:   0.08, prob_loss:   2.37, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(5.859969, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   18/125, lr:0.000003, giou_loss:   2.86, conf_loss:   0.08, prob_loss:   2.92, total_loss:   5.86\n",
            " Total Loss  tf.Tensor(4.8365693, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   19/125, lr:0.000003, giou_loss:   2.31, conf_loss:   0.48, prob_loss:   2.05, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(5.1878757, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   20/125, lr:0.000003, giou_loss:   2.32, conf_loss:   0.07, prob_loss:   2.79, total_loss:   5.19\n",
            " Total Loss  tf.Tensor(3.7999833, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   21/125, lr:0.000003, giou_loss:   1.77, conf_loss:   0.02, prob_loss:   2.00, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(4.827237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   22/125, lr:0.000003, giou_loss:   2.27, conf_loss:   0.03, prob_loss:   2.53, total_loss:   4.83\n",
            " Total Loss  tf.Tensor(4.2072163, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   23/125, lr:0.000003, giou_loss:   1.87, conf_loss:   0.40, prob_loss:   1.93, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.093771, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   24/125, lr:0.000003, giou_loss:   1.85, conf_loss:   0.06, prob_loss:   2.19, total_loss:   4.09\n",
            " Total Loss  tf.Tensor(3.8948221, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   25/125, lr:0.000003, giou_loss:   1.81, conf_loss:   0.03, prob_loss:   2.05, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(5.3372846, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   26/125, lr:0.000003, giou_loss:   2.52, conf_loss:   0.07, prob_loss:   2.75, total_loss:   5.34\n",
            " Total Loss  tf.Tensor(5.329539, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   27/125, lr:0.000003, giou_loss:   2.34, conf_loss:   0.03, prob_loss:   2.96, total_loss:   5.33\n",
            " Total Loss  tf.Tensor(3.9658108, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   28/125, lr:0.000003, giou_loss:   1.87, conf_loss:   0.14, prob_loss:   1.95, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(3.1081572, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   29/125, lr:0.000003, giou_loss:   1.52, conf_loss:   0.03, prob_loss:   1.57, total_loss:   3.11\n",
            " Total Loss  tf.Tensor(4.406289, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   30/125, lr:0.000003, giou_loss:   2.12, conf_loss:   0.06, prob_loss:   2.23, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(3.4375567, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   31/125, lr:0.000003, giou_loss:   1.42, conf_loss:   0.29, prob_loss:   1.74, total_loss:   3.44\n",
            " Total Loss  tf.Tensor(5.5865436, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   32/125, lr:0.000003, giou_loss:   2.72, conf_loss:   0.03, prob_loss:   2.84, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(4.387474, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   33/125, lr:0.000003, giou_loss:   1.89, conf_loss:   0.09, prob_loss:   2.41, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(4.5602064, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   34/125, lr:0.000003, giou_loss:   1.87, conf_loss:   0.18, prob_loss:   2.51, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.687825, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   35/125, lr:0.000003, giou_loss:   2.17, conf_loss:   0.12, prob_loss:   2.40, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(4.4796424, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   36/125, lr:0.000003, giou_loss:   2.11, conf_loss:   0.30, prob_loss:   2.06, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(3.3208623, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   37/125, lr:0.000003, giou_loss:   1.43, conf_loss:   0.01, prob_loss:   1.88, total_loss:   3.32\n",
            " Total Loss  tf.Tensor(4.36057, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   38/125, lr:0.000002, giou_loss:   2.01, conf_loss:   0.09, prob_loss:   2.26, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.8656497, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   39/125, lr:0.000002, giou_loss:   2.30, conf_loss:   0.04, prob_loss:   2.52, total_loss:   4.87\n",
            " Total Loss  tf.Tensor(4.5844774, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   40/125, lr:0.000002, giou_loss:   2.19, conf_loss:   0.03, prob_loss:   2.37, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.655785, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   41/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.05, prob_loss:   2.54, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.4664974, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   42/125, lr:0.000002, giou_loss:   2.21, conf_loss:   0.03, prob_loss:   2.23, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(4.5275974, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   43/125, lr:0.000002, giou_loss:   2.17, conf_loss:   0.02, prob_loss:   2.33, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(5.1533656, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   44/125, lr:0.000002, giou_loss:   2.42, conf_loss:   0.03, prob_loss:   2.70, total_loss:   5.15\n",
            " Total Loss  tf.Tensor(4.226712, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   45/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.05, prob_loss:   2.19, total_loss:   4.23\n",
            " Total Loss  tf.Tensor(5.265909, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   46/125, lr:0.000002, giou_loss:   2.43, conf_loss:   0.06, prob_loss:   2.78, total_loss:   5.27\n",
            " Total Loss  tf.Tensor(2.980133, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   47/125, lr:0.000002, giou_loss:   1.42, conf_loss:   0.04, prob_loss:   1.51, total_loss:   2.98\n",
            " Total Loss  tf.Tensor(4.2190533, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   48/125, lr:0.000002, giou_loss:   1.95, conf_loss:   0.03, prob_loss:   2.25, total_loss:   4.22\n",
            " Total Loss  tf.Tensor(3.577671, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   49/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.08, prob_loss:   1.70, total_loss:   3.58\n",
            " Total Loss  tf.Tensor(3.490843, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   50/125, lr:0.000002, giou_loss:   1.49, conf_loss:   0.01, prob_loss:   1.99, total_loss:   3.49\n",
            " Total Loss  tf.Tensor(4.0631943, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   51/125, lr:0.000002, giou_loss:   1.85, conf_loss:   0.07, prob_loss:   2.14, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(4.1450934, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   52/125, lr:0.000002, giou_loss:   1.92, conf_loss:   0.08, prob_loss:   2.14, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(5.184746, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   53/125, lr:0.000002, giou_loss:   2.40, conf_loss:   0.15, prob_loss:   2.64, total_loss:   5.18\n",
            " Total Loss  tf.Tensor(3.7700562, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   54/125, lr:0.000002, giou_loss:   1.92, conf_loss:   0.03, prob_loss:   1.82, total_loss:   3.77\n",
            " Total Loss  tf.Tensor(5.8882856, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   55/125, lr:0.000002, giou_loss:   2.67, conf_loss:   0.19, prob_loss:   3.03, total_loss:   5.89\n",
            " Total Loss  tf.Tensor(5.972954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   56/125, lr:0.000002, giou_loss:   2.79, conf_loss:   0.08, prob_loss:   3.11, total_loss:   5.97\n",
            " Total Loss  tf.Tensor(4.9752417, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   57/125, lr:0.000002, giou_loss:   2.30, conf_loss:   0.13, prob_loss:   2.54, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(5.362001, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   58/125, lr:0.000002, giou_loss:   2.52, conf_loss:   0.12, prob_loss:   2.72, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(4.0773344, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   59/125, lr:0.000002, giou_loss:   1.86, conf_loss:   0.03, prob_loss:   2.19, total_loss:   4.08\n",
            " Total Loss  tf.Tensor(4.6941147, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   60/125, lr:0.000002, giou_loss:   2.23, conf_loss:   0.05, prob_loss:   2.41, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(4.8756723, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   61/125, lr:0.000002, giou_loss:   2.37, conf_loss:   0.39, prob_loss:   2.11, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(4.6873884, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   62/125, lr:0.000002, giou_loss:   2.48, conf_loss:   0.06, prob_loss:   2.16, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(5.0884695, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   63/125, lr:0.000002, giou_loss:   2.30, conf_loss:   0.04, prob_loss:   2.75, total_loss:   5.09\n",
            " Total Loss  tf.Tensor(3.3501935, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   64/125, lr:0.000002, giou_loss:   1.50, conf_loss:   0.13, prob_loss:   1.72, total_loss:   3.35\n",
            " Total Loss  tf.Tensor(3.6380706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   65/125, lr:0.000002, giou_loss:   1.75, conf_loss:   0.07, prob_loss:   1.82, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(3.8590534, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   66/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.13, prob_loss:   1.93, total_loss:   3.86\n",
            " Total Loss  tf.Tensor(3.7438185, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   67/125, lr:0.000002, giou_loss:   1.61, conf_loss:   0.30, prob_loss:   1.83, total_loss:   3.74\n",
            " Total Loss  tf.Tensor(4.4775667, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   68/125, lr:0.000002, giou_loss:   2.18, conf_loss:   0.01, prob_loss:   2.29, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(5.8158073, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   69/125, lr:0.000002, giou_loss:   2.68, conf_loss:   0.31, prob_loss:   2.83, total_loss:   5.82\n",
            " Total Loss  tf.Tensor(3.9630346, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   70/125, lr:0.000002, giou_loss:   1.89, conf_loss:   0.03, prob_loss:   2.04, total_loss:   3.96\n",
            " Total Loss  tf.Tensor(4.863924, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   71/125, lr:0.000002, giou_loss:   2.19, conf_loss:   0.11, prob_loss:   2.56, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(5.3570967, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   72/125, lr:0.000002, giou_loss:   2.31, conf_loss:   0.07, prob_loss:   2.97, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(3.2754285, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   73/125, lr:0.000002, giou_loss:   1.50, conf_loss:   0.06, prob_loss:   1.72, total_loss:   3.28\n",
            " Total Loss  tf.Tensor(4.8308625, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   74/125, lr:0.000002, giou_loss:   2.19, conf_loss:   0.27, prob_loss:   2.37, total_loss:   4.83\n",
            " Total Loss  tf.Tensor(4.0206428, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   75/125, lr:0.000002, giou_loss:   1.75, conf_loss:   0.02, prob_loss:   2.25, total_loss:   4.02\n",
            " Total Loss  tf.Tensor(4.4210553, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   76/125, lr:0.000002, giou_loss:   1.90, conf_loss:   0.03, prob_loss:   2.49, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(5.506978, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   77/125, lr:0.000002, giou_loss:   2.51, conf_loss:   0.13, prob_loss:   2.87, total_loss:   5.51\n",
            " Total Loss  tf.Tensor(4.5986996, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   78/125, lr:0.000002, giou_loss:   2.20, conf_loss:   0.02, prob_loss:   2.37, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(4.1066403, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   79/125, lr:0.000002, giou_loss:   2.15, conf_loss:   0.06, prob_loss:   1.89, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(5.012376, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   80/125, lr:0.000002, giou_loss:   2.45, conf_loss:   0.16, prob_loss:   2.41, total_loss:   5.01\n",
            " Total Loss  tf.Tensor(4.132321, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   81/125, lr:0.000002, giou_loss:   1.75, conf_loss:   0.11, prob_loss:   2.27, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(4.2049923, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   82/125, lr:0.000002, giou_loss:   2.01, conf_loss:   0.02, prob_loss:   2.17, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(4.473445, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   83/125, lr:0.000002, giou_loss:   1.90, conf_loss:   0.23, prob_loss:   2.34, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(4.3915877, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   84/125, lr:0.000002, giou_loss:   2.04, conf_loss:   0.09, prob_loss:   2.26, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(4.4004, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   85/125, lr:0.000002, giou_loss:   1.93, conf_loss:   0.18, prob_loss:   2.29, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(4.533714, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   86/125, lr:0.000002, giou_loss:   2.22, conf_loss:   0.10, prob_loss:   2.21, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(5.42047, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   87/125, lr:0.000002, giou_loss:   2.54, conf_loss:   0.34, prob_loss:   2.54, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(4.0062704, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   88/125, lr:0.000002, giou_loss:   2.03, conf_loss:   0.06, prob_loss:   1.92, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(5.3335104, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   89/125, lr:0.000002, giou_loss:   2.69, conf_loss:   0.32, prob_loss:   2.32, total_loss:   5.33\n",
            " Total Loss  tf.Tensor(4.199506, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   90/125, lr:0.000002, giou_loss:   1.88, conf_loss:   0.03, prob_loss:   2.28, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(3.9714732, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   91/125, lr:0.000002, giou_loss:   1.73, conf_loss:   0.07, prob_loss:   2.17, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(4.2903867, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   92/125, lr:0.000002, giou_loss:   1.83, conf_loss:   0.06, prob_loss:   2.40, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.9353604, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   93/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.07, prob_loss:   2.80, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(5.666914, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   94/125, lr:0.000002, giou_loss:   2.51, conf_loss:   0.26, prob_loss:   2.90, total_loss:   5.67\n",
            " Total Loss  tf.Tensor(3.8516312, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   95/125, lr:0.000002, giou_loss:   1.66, conf_loss:   0.33, prob_loss:   1.86, total_loss:   3.85\n",
            " Total Loss  tf.Tensor(4.6875134, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   96/125, lr:0.000002, giou_loss:   2.38, conf_loss:   0.10, prob_loss:   2.21, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(5.2486553, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   97/125, lr:0.000002, giou_loss:   2.59, conf_loss:   0.11, prob_loss:   2.55, total_loss:   5.25\n",
            " Total Loss  tf.Tensor(5.548525, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   98/125, lr:0.000002, giou_loss:   2.71, conf_loss:   0.07, prob_loss:   2.76, total_loss:   5.55\n",
            " Total Loss  tf.Tensor(4.6388373, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:   99/125, lr:0.000002, giou_loss:   2.35, conf_loss:   0.08, prob_loss:   2.21, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.474683, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  100/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.06, prob_loss:   2.34, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(4.0094633, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  101/125, lr:0.000002, giou_loss:   1.75, conf_loss:   0.09, prob_loss:   2.17, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(5.4021225, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  102/125, lr:0.000002, giou_loss:   2.53, conf_loss:   0.11, prob_loss:   2.77, total_loss:   5.40\n",
            " Total Loss  tf.Tensor(3.9419549, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  103/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.03, prob_loss:   2.19, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(4.202335, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  104/125, lr:0.000002, giou_loss:   1.83, conf_loss:   0.10, prob_loss:   2.27, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(4.61077, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  105/125, lr:0.000002, giou_loss:   2.13, conf_loss:   0.13, prob_loss:   2.35, total_loss:   4.61\n",
            " Total Loss  tf.Tensor(4.694769, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  106/125, lr:0.000002, giou_loss:   2.08, conf_loss:   0.13, prob_loss:   2.49, total_loss:   4.69\n",
            " Total Loss  tf.Tensor(3.9264793, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  107/125, lr:0.000002, giou_loss:   1.77, conf_loss:   0.10, prob_loss:   2.06, total_loss:   3.93\n",
            " Total Loss  tf.Tensor(4.906745, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  108/125, lr:0.000002, giou_loss:   2.42, conf_loss:   0.04, prob_loss:   2.44, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(4.9624715, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  109/125, lr:0.000002, giou_loss:   2.43, conf_loss:   0.05, prob_loss:   2.48, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(3.899759, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  110/125, lr:0.000002, giou_loss:   1.63, conf_loss:   0.05, prob_loss:   2.22, total_loss:   3.90\n",
            " Total Loss  tf.Tensor(5.3568544, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  111/125, lr:0.000002, giou_loss:   2.57, conf_loss:   0.13, prob_loss:   2.66, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(4.1764584, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  112/125, lr:0.000002, giou_loss:   1.92, conf_loss:   0.07, prob_loss:   2.19, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(5.4865065, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  113/125, lr:0.000002, giou_loss:   2.67, conf_loss:   0.15, prob_loss:   2.67, total_loss:   5.49\n",
            " Total Loss  tf.Tensor(5.6218386, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  114/125, lr:0.000002, giou_loss:   2.76, conf_loss:   0.12, prob_loss:   2.74, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(4.244747, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  115/125, lr:0.000002, giou_loss:   2.00, conf_loss:   0.03, prob_loss:   2.21, total_loss:   4.24\n",
            " Total Loss  tf.Tensor(5.053404, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  116/125, lr:0.000002, giou_loss:   2.25, conf_loss:   0.08, prob_loss:   2.73, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(6.9010506, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  117/125, lr:0.000002, giou_loss:   3.36, conf_loss:   0.46, prob_loss:   3.08, total_loss:   6.90\n",
            " Total Loss  tf.Tensor(4.2043223, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  118/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.05, prob_loss:   2.13, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(5.4188232, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  119/125, lr:0.000002, giou_loss:   2.50, conf_loss:   0.10, prob_loss:   2.83, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(3.1805148, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  120/125, lr:0.000002, giou_loss:   1.66, conf_loss:   0.10, prob_loss:   1.42, total_loss:   3.18\n",
            " Total Loss  tf.Tensor(5.7817707, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  121/125, lr:0.000002, giou_loss:   2.78, conf_loss:   0.03, prob_loss:   2.97, total_loss:   5.78\n",
            " Total Loss  tf.Tensor(6.1584263, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  122/125, lr:0.000002, giou_loss:   2.69, conf_loss:   0.17, prob_loss:   3.29, total_loss:   6.16\n",
            " Total Loss  tf.Tensor(3.4544, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  123/125, lr:0.000002, giou_loss:   1.50, conf_loss:   0.12, prob_loss:   1.83, total_loss:   3.45\n",
            " Total Loss  tf.Tensor(4.696137, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:  124/125, lr:0.000002, giou_loss:   2.12, conf_loss:   0.15, prob_loss:   2.42, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(4.7522383, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    0/125, lr:0.000002, giou_loss:   2.30, conf_loss:   0.04, prob_loss:   2.41, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(5.0274, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:92 step:    1/125, lr:0.000002, giou_loss:   2.39, conf_loss:   0.13, prob_loss:   2.51, total_loss:   5.03\n",
            "\n",
            "\n",
            "giou_val_loss:   2.22, conf_val_loss:   0.62, prob_val_loss:   5.22, total_val_loss:   8.07\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.9310374, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    2/125, lr:0.000002, giou_loss:   2.46, conf_loss:   0.07, prob_loss:   2.41, total_loss:   4.93\n",
            " Total Loss  tf.Tensor(2.7992635, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    3/125, lr:0.000002, giou_loss:   1.43, conf_loss:   0.06, prob_loss:   1.31, total_loss:   2.80\n",
            " Total Loss  tf.Tensor(5.598373, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    4/125, lr:0.000002, giou_loss:   2.68, conf_loss:   0.11, prob_loss:   2.81, total_loss:   5.60\n",
            " Total Loss  tf.Tensor(5.6110086, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    5/125, lr:0.000002, giou_loss:   2.64, conf_loss:   0.10, prob_loss:   2.87, total_loss:   5.61\n",
            " Total Loss  tf.Tensor(5.141291, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    6/125, lr:0.000002, giou_loss:   2.32, conf_loss:   0.04, prob_loss:   2.78, total_loss:   5.14\n",
            " Total Loss  tf.Tensor(4.5766954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    7/125, lr:0.000002, giou_loss:   2.21, conf_loss:   0.12, prob_loss:   2.25, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.501749, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    8/125, lr:0.000002, giou_loss:   2.41, conf_loss:   0.04, prob_loss:   2.05, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.5625196, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    9/125, lr:0.000002, giou_loss:   2.10, conf_loss:   0.09, prob_loss:   2.37, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(5.1136465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   10/125, lr:0.000002, giou_loss:   2.50, conf_loss:   0.10, prob_loss:   2.51, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(3.8927124, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   11/125, lr:0.000002, giou_loss:   1.81, conf_loss:   0.03, prob_loss:   2.06, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(3.747417, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   12/125, lr:0.000002, giou_loss:   1.77, conf_loss:   0.02, prob_loss:   1.96, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(4.855466, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   13/125, lr:0.000002, giou_loss:   2.37, conf_loss:   0.03, prob_loss:   2.46, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(4.996924, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   14/125, lr:0.000002, giou_loss:   2.36, conf_loss:   0.12, prob_loss:   2.52, total_loss:   5.00\n",
            " Total Loss  tf.Tensor(6.8155804, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   15/125, lr:0.000002, giou_loss:   3.33, conf_loss:   0.13, prob_loss:   3.36, total_loss:   6.82\n",
            " Total Loss  tf.Tensor(3.8407304, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   16/125, lr:0.000002, giou_loss:   1.70, conf_loss:   0.14, prob_loss:   2.00, total_loss:   3.84\n",
            " Total Loss  tf.Tensor(4.270916, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   17/125, lr:0.000002, giou_loss:   2.04, conf_loss:   0.04, prob_loss:   2.19, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(6.34066, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   18/125, lr:0.000002, giou_loss:   3.10, conf_loss:   0.09, prob_loss:   3.15, total_loss:   6.34\n",
            " Total Loss  tf.Tensor(4.3595076, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   19/125, lr:0.000002, giou_loss:   2.15, conf_loss:   0.03, prob_loss:   2.18, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(6.6076837, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   20/125, lr:0.000002, giou_loss:   2.67, conf_loss:   0.18, prob_loss:   3.76, total_loss:   6.61\n",
            " Total Loss  tf.Tensor(4.6187754, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   21/125, lr:0.000002, giou_loss:   1.93, conf_loss:   0.09, prob_loss:   2.60, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(4.9110174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   22/125, lr:0.000002, giou_loss:   2.33, conf_loss:   0.02, prob_loss:   2.56, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(4.9003773, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   23/125, lr:0.000002, giou_loss:   2.28, conf_loss:   0.13, prob_loss:   2.48, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(4.5059834, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   24/125, lr:0.000002, giou_loss:   2.12, conf_loss:   0.02, prob_loss:   2.36, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.733617, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   25/125, lr:0.000002, giou_loss:   1.94, conf_loss:   0.57, prob_loss:   2.22, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(3.4460266, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   26/125, lr:0.000002, giou_loss:   1.58, conf_loss:   0.03, prob_loss:   1.84, total_loss:   3.45\n",
            " Total Loss  tf.Tensor(4.588023, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   27/125, lr:0.000002, giou_loss:   2.05, conf_loss:   0.21, prob_loss:   2.33, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(4.844016, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   28/125, lr:0.000002, giou_loss:   2.24, conf_loss:   0.20, prob_loss:   2.40, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.270502, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   29/125, lr:0.000002, giou_loss:   1.82, conf_loss:   0.06, prob_loss:   2.39, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(4.206127, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   30/125, lr:0.000002, giou_loss:   1.78, conf_loss:   0.10, prob_loss:   2.32, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.071679, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   31/125, lr:0.000002, giou_loss:   2.03, conf_loss:   0.05, prob_loss:   2.00, total_loss:   4.07\n",
            " Total Loss  tf.Tensor(3.4130516, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   32/125, lr:0.000002, giou_loss:   1.66, conf_loss:   0.05, prob_loss:   1.70, total_loss:   3.41\n",
            " Total Loss  tf.Tensor(4.5024095, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   33/125, lr:0.000002, giou_loss:   2.17, conf_loss:   0.03, prob_loss:   2.30, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.297791, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   34/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.11, prob_loss:   2.46, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.36193, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   35/125, lr:0.000002, giou_loss:   1.97, conf_loss:   0.13, prob_loss:   2.26, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(3.982216, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   36/125, lr:0.000002, giou_loss:   1.89, conf_loss:   0.03, prob_loss:   2.06, total_loss:   3.98\n",
            " Total Loss  tf.Tensor(4.308693, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   37/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.16, prob_loss:   1.99, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.6037087, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   38/125, lr:0.000002, giou_loss:   1.94, conf_loss:   0.01, prob_loss:   2.66, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(6.2796936, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   39/125, lr:0.000002, giou_loss:   2.71, conf_loss:   0.16, prob_loss:   3.41, total_loss:   6.28\n",
            " Total Loss  tf.Tensor(4.445945, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   40/125, lr:0.000002, giou_loss:   2.06, conf_loss:   0.20, prob_loss:   2.19, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(3.762296, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   41/125, lr:0.000002, giou_loss:   1.67, conf_loss:   0.05, prob_loss:   2.04, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(3.3568726, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   42/125, lr:0.000002, giou_loss:   1.52, conf_loss:   0.04, prob_loss:   1.80, total_loss:   3.36\n",
            " Total Loss  tf.Tensor(4.665163, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   43/125, lr:0.000002, giou_loss:   2.23, conf_loss:   0.04, prob_loss:   2.39, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.711062, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   44/125, lr:0.000002, giou_loss:   2.32, conf_loss:   0.23, prob_loss:   3.16, total_loss:   5.71\n",
            " Total Loss  tf.Tensor(4.655208, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   45/125, lr:0.000002, giou_loss:   2.28, conf_loss:   0.03, prob_loss:   2.35, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.3750057, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   46/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.04, prob_loss:   2.35, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(4.2937136, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   47/125, lr:0.000002, giou_loss:   1.82, conf_loss:   0.03, prob_loss:   2.45, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(3.7211852, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   48/125, lr:0.000002, giou_loss:   1.68, conf_loss:   0.05, prob_loss:   1.99, total_loss:   3.72\n",
            " Total Loss  tf.Tensor(4.6199102, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   49/125, lr:0.000002, giou_loss:   2.06, conf_loss:   0.10, prob_loss:   2.45, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(4.446671, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   50/125, lr:0.000002, giou_loss:   2.15, conf_loss:   0.02, prob_loss:   2.28, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(5.658358, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   51/125, lr:0.000002, giou_loss:   2.67, conf_loss:   0.25, prob_loss:   2.74, total_loss:   5.66\n",
            " Total Loss  tf.Tensor(3.521101, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   52/125, lr:0.000002, giou_loss:   1.64, conf_loss:   0.06, prob_loss:   1.82, total_loss:   3.52\n",
            " Total Loss  tf.Tensor(4.526529, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   53/125, lr:0.000002, giou_loss:   2.04, conf_loss:   0.03, prob_loss:   2.45, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(4.9856567, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   54/125, lr:0.000002, giou_loss:   2.36, conf_loss:   0.17, prob_loss:   2.45, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(3.41989, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   55/125, lr:0.000002, giou_loss:   1.73, conf_loss:   0.03, prob_loss:   1.66, total_loss:   3.42\n",
            " Total Loss  tf.Tensor(4.396281, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   56/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.12, prob_loss:   2.25, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(4.357268, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   57/125, lr:0.000002, giou_loss:   1.83, conf_loss:   0.03, prob_loss:   2.49, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.4547424, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   58/125, lr:0.000002, giou_loss:   1.88, conf_loss:   0.37, prob_loss:   2.21, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.377815, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   59/125, lr:0.000002, giou_loss:   1.95, conf_loss:   0.08, prob_loss:   2.35, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(5.6919413, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   60/125, lr:0.000002, giou_loss:   2.86, conf_loss:   0.03, prob_loss:   2.80, total_loss:   5.69\n",
            " Total Loss  tf.Tensor(5.0722713, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   61/125, lr:0.000002, giou_loss:   2.50, conf_loss:   0.08, prob_loss:   2.50, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(4.8313303, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   62/125, lr:0.000002, giou_loss:   2.13, conf_loss:   0.25, prob_loss:   2.45, total_loss:   4.83\n",
            " Total Loss  tf.Tensor(4.296627, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   63/125, lr:0.000002, giou_loss:   1.74, conf_loss:   0.03, prob_loss:   2.52, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.146985, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   64/125, lr:0.000002, giou_loss:   1.77, conf_loss:   0.02, prob_loss:   2.36, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(3.9250004, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   65/125, lr:0.000002, giou_loss:   1.89, conf_loss:   0.13, prob_loss:   1.90, total_loss:   3.93\n",
            " Total Loss  tf.Tensor(4.92245, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   66/125, lr:0.000002, giou_loss:   2.24, conf_loss:   0.31, prob_loss:   2.37, total_loss:   4.92\n",
            " Total Loss  tf.Tensor(3.7513623, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   67/125, lr:0.000002, giou_loss:   1.80, conf_loss:   0.06, prob_loss:   1.89, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(4.4102983, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   68/125, lr:0.000002, giou_loss:   2.14, conf_loss:   0.15, prob_loss:   2.12, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(4.049572, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   69/125, lr:0.000002, giou_loss:   1.86, conf_loss:   0.12, prob_loss:   2.06, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(3.4328496, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   70/125, lr:0.000002, giou_loss:   1.53, conf_loss:   0.01, prob_loss:   1.89, total_loss:   3.43\n",
            " Total Loss  tf.Tensor(5.5333776, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   71/125, lr:0.000002, giou_loss:   2.88, conf_loss:   0.05, prob_loss:   2.61, total_loss:   5.53\n",
            " Total Loss  tf.Tensor(2.921214, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   72/125, lr:0.000002, giou_loss:   1.31, conf_loss:   0.02, prob_loss:   1.59, total_loss:   2.92\n",
            " Total Loss  tf.Tensor(3.265365, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   73/125, lr:0.000002, giou_loss:   1.53, conf_loss:   0.04, prob_loss:   1.69, total_loss:   3.27\n",
            " Total Loss  tf.Tensor(5.075138, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   74/125, lr:0.000002, giou_loss:   2.17, conf_loss:   0.40, prob_loss:   2.50, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(4.181738, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   75/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.07, prob_loss:   2.12, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(5.171516, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   76/125, lr:0.000002, giou_loss:   2.55, conf_loss:   0.12, prob_loss:   2.50, total_loss:   5.17\n",
            " Total Loss  tf.Tensor(5.073851, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   77/125, lr:0.000002, giou_loss:   2.33, conf_loss:   0.02, prob_loss:   2.72, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(4.4139776, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   78/125, lr:0.000002, giou_loss:   2.15, conf_loss:   0.02, prob_loss:   2.24, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(5.1728535, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   79/125, lr:0.000002, giou_loss:   2.28, conf_loss:   0.06, prob_loss:   2.83, total_loss:   5.17\n",
            " Total Loss  tf.Tensor(3.3007874, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   80/125, lr:0.000002, giou_loss:   1.59, conf_loss:   0.04, prob_loss:   1.68, total_loss:   3.30\n",
            " Total Loss  tf.Tensor(5.4653616, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   81/125, lr:0.000002, giou_loss:   2.44, conf_loss:   0.05, prob_loss:   2.97, total_loss:   5.47\n",
            " Total Loss  tf.Tensor(4.728721, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   82/125, lr:0.000002, giou_loss:   1.91, conf_loss:   0.09, prob_loss:   2.73, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(3.988525, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   83/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.03, prob_loss:   2.17, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(4.957639, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   84/125, lr:0.000002, giou_loss:   2.42, conf_loss:   0.23, prob_loss:   2.30, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(4.037328, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   85/125, lr:0.000002, giou_loss:   2.12, conf_loss:   0.06, prob_loss:   1.85, total_loss:   4.04\n",
            " Total Loss  tf.Tensor(3.475632, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   86/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.05, prob_loss:   1.70, total_loss:   3.48\n",
            " Total Loss  tf.Tensor(5.187646, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   87/125, lr:0.000002, giou_loss:   2.40, conf_loss:   0.03, prob_loss:   2.76, total_loss:   5.19\n",
            " Total Loss  tf.Tensor(3.5233738, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   88/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.02, prob_loss:   1.79, total_loss:   3.52\n",
            " Total Loss  tf.Tensor(4.4624424, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   89/125, lr:0.000002, giou_loss:   2.12, conf_loss:   0.07, prob_loss:   2.27, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(5.3720217, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   90/125, lr:0.000002, giou_loss:   2.57, conf_loss:   0.04, prob_loss:   2.77, total_loss:   5.37\n",
            " Total Loss  tf.Tensor(4.1568723, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   91/125, lr:0.000002, giou_loss:   1.96, conf_loss:   0.03, prob_loss:   2.17, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(4.6209908, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   92/125, lr:0.000002, giou_loss:   1.96, conf_loss:   0.16, prob_loss:   2.50, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(5.3861647, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   93/125, lr:0.000002, giou_loss:   2.43, conf_loss:   0.04, prob_loss:   2.92, total_loss:   5.39\n",
            " Total Loss  tf.Tensor(5.3963823, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   94/125, lr:0.000002, giou_loss:   2.68, conf_loss:   0.19, prob_loss:   2.52, total_loss:   5.40\n",
            " Total Loss  tf.Tensor(4.496801, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   95/125, lr:0.000002, giou_loss:   2.09, conf_loss:   0.04, prob_loss:   2.36, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.078292, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   96/125, lr:0.000002, giou_loss:   1.91, conf_loss:   0.01, prob_loss:   2.16, total_loss:   4.08\n",
            " Total Loss  tf.Tensor(4.4168096, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   97/125, lr:0.000002, giou_loss:   2.34, conf_loss:   0.02, prob_loss:   2.06, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(4.381654, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   98/125, lr:0.000002, giou_loss:   2.09, conf_loss:   0.05, prob_loss:   2.24, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(4.914992, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:   99/125, lr:0.000002, giou_loss:   2.18, conf_loss:   0.10, prob_loss:   2.63, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(3.636797, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  100/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.08, prob_loss:   1.77, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(3.653784, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  101/125, lr:0.000002, giou_loss:   1.59, conf_loss:   0.03, prob_loss:   2.04, total_loss:   3.65\n",
            " Total Loss  tf.Tensor(4.515168, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  102/125, lr:0.000002, giou_loss:   2.05, conf_loss:   0.05, prob_loss:   2.42, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(4.802987, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  103/125, lr:0.000002, giou_loss:   2.28, conf_loss:   0.03, prob_loss:   2.50, total_loss:   4.80\n",
            " Total Loss  tf.Tensor(7.1714325, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  104/125, lr:0.000002, giou_loss:   2.68, conf_loss:   0.39, prob_loss:   4.10, total_loss:   7.17\n",
            " Total Loss  tf.Tensor(4.314994, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  105/125, lr:0.000002, giou_loss:   2.01, conf_loss:   0.04, prob_loss:   2.26, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.398554, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  106/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.05, prob_loss:   2.34, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(4.6397343, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  107/125, lr:0.000002, giou_loss:   2.10, conf_loss:   0.05, prob_loss:   2.49, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(3.655758, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  108/125, lr:0.000002, giou_loss:   1.65, conf_loss:   0.03, prob_loss:   1.98, total_loss:   3.66\n",
            " Total Loss  tf.Tensor(5.325556, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  109/125, lr:0.000002, giou_loss:   2.50, conf_loss:   0.25, prob_loss:   2.58, total_loss:   5.33\n",
            " Total Loss  tf.Tensor(5.538223, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  110/125, lr:0.000002, giou_loss:   2.82, conf_loss:   0.22, prob_loss:   2.50, total_loss:   5.54\n",
            " Total Loss  tf.Tensor(6.25723, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  111/125, lr:0.000002, giou_loss:   2.98, conf_loss:   0.13, prob_loss:   3.15, total_loss:   6.26\n",
            " Total Loss  tf.Tensor(4.7278595, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  112/125, lr:0.000002, giou_loss:   1.95, conf_loss:   0.03, prob_loss:   2.74, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(4.6096926, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  113/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.14, prob_loss:   2.31, total_loss:   4.61\n",
            " Total Loss  tf.Tensor(4.266085, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  114/125, lr:0.000002, giou_loss:   2.19, conf_loss:   0.02, prob_loss:   2.06, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(5.0650015, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  115/125, lr:0.000002, giou_loss:   2.34, conf_loss:   0.04, prob_loss:   2.69, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(5.1697416, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  116/125, lr:0.000002, giou_loss:   2.46, conf_loss:   0.02, prob_loss:   2.68, total_loss:   5.17\n",
            " Total Loss  tf.Tensor(4.7443914, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  117/125, lr:0.000002, giou_loss:   2.18, conf_loss:   0.18, prob_loss:   2.39, total_loss:   4.74\n",
            " Total Loss  tf.Tensor(4.172078, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  118/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.08, prob_loss:   2.10, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.1017923, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  119/125, lr:0.000002, giou_loss:   1.91, conf_loss:   0.05, prob_loss:   2.14, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(5.3842287, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  120/125, lr:0.000002, giou_loss:   2.41, conf_loss:   0.19, prob_loss:   2.78, total_loss:   5.38\n",
            " Total Loss  tf.Tensor(3.236507, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  121/125, lr:0.000002, giou_loss:   1.40, conf_loss:   0.05, prob_loss:   1.79, total_loss:   3.24\n",
            " Total Loss  tf.Tensor(3.3274627, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  122/125, lr:0.000002, giou_loss:   1.37, conf_loss:   0.01, prob_loss:   1.95, total_loss:   3.33\n",
            " Total Loss  tf.Tensor(4.7677155, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  123/125, lr:0.000002, giou_loss:   2.17, conf_loss:   0.03, prob_loss:   2.57, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(4.988677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:  124/125, lr:0.000002, giou_loss:   2.27, conf_loss:   0.05, prob_loss:   2.67, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(4.575232, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    0/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.12, prob_loss:   2.30, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.9921436, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:93 step:    1/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.11, prob_loss:   2.09, total_loss:   3.99\n",
            "\n",
            "\n",
            "giou_val_loss:   2.25, conf_val_loss:   0.61, prob_val_loss:   5.49, total_val_loss:   8.35\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.320816, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    2/125, lr:0.000002, giou_loss:   1.89, conf_loss:   0.08, prob_loss:   2.35, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(5.551049, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    3/125, lr:0.000002, giou_loss:   2.49, conf_loss:   0.33, prob_loss:   2.73, total_loss:   5.55\n",
            " Total Loss  tf.Tensor(4.938298, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    4/125, lr:0.000002, giou_loss:   2.61, conf_loss:   0.23, prob_loss:   2.10, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(3.5487957, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    5/125, lr:0.000002, giou_loss:   1.74, conf_loss:   0.12, prob_loss:   1.68, total_loss:   3.55\n",
            " Total Loss  tf.Tensor(4.102128, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    6/125, lr:0.000002, giou_loss:   1.81, conf_loss:   0.03, prob_loss:   2.26, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(3.2619452, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    7/125, lr:0.000002, giou_loss:   1.46, conf_loss:   0.02, prob_loss:   1.78, total_loss:   3.26\n",
            " Total Loss  tf.Tensor(3.8257174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    8/125, lr:0.000002, giou_loss:   1.80, conf_loss:   0.04, prob_loss:   1.98, total_loss:   3.83\n",
            " Total Loss  tf.Tensor(4.4095993, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    9/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.03, prob_loss:   2.36, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(2.3621035, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   10/125, lr:0.000002, giou_loss:   1.03, conf_loss:   0.02, prob_loss:   1.31, total_loss:   2.36\n",
            " Total Loss  tf.Tensor(5.307109, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   11/125, lr:0.000002, giou_loss:   2.60, conf_loss:   0.28, prob_loss:   2.42, total_loss:   5.31\n",
            " Total Loss  tf.Tensor(4.553906, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   12/125, lr:0.000002, giou_loss:   2.05, conf_loss:   0.22, prob_loss:   2.29, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(3.6506925, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   13/125, lr:0.000002, giou_loss:   1.60, conf_loss:   0.03, prob_loss:   2.03, total_loss:   3.65\n",
            " Total Loss  tf.Tensor(5.175305, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   14/125, lr:0.000002, giou_loss:   2.49, conf_loss:   0.10, prob_loss:   2.60, total_loss:   5.18\n",
            " Total Loss  tf.Tensor(4.1049023, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   15/125, lr:0.000002, giou_loss:   1.75, conf_loss:   0.06, prob_loss:   2.29, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(5.2739515, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   16/125, lr:0.000002, giou_loss:   2.40, conf_loss:   0.21, prob_loss:   2.66, total_loss:   5.27\n",
            " Total Loss  tf.Tensor(5.537857, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   17/125, lr:0.000002, giou_loss:   2.54, conf_loss:   0.19, prob_loss:   2.81, total_loss:   5.54\n",
            " Total Loss  tf.Tensor(3.2875552, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   18/125, lr:0.000002, giou_loss:   1.76, conf_loss:   0.20, prob_loss:   1.33, total_loss:   3.29\n",
            " Total Loss  tf.Tensor(4.893479, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   19/125, lr:0.000002, giou_loss:   2.34, conf_loss:   0.05, prob_loss:   2.50, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(3.1288471, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   20/125, lr:0.000002, giou_loss:   1.45, conf_loss:   0.03, prob_loss:   1.65, total_loss:   3.13\n",
            " Total Loss  tf.Tensor(5.2812495, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   21/125, lr:0.000002, giou_loss:   2.47, conf_loss:   0.25, prob_loss:   2.56, total_loss:   5.28\n",
            " Total Loss  tf.Tensor(4.411509, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   22/125, lr:0.000002, giou_loss:   2.06, conf_loss:   0.04, prob_loss:   2.31, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(3.3675613, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   23/125, lr:0.000002, giou_loss:   1.53, conf_loss:   0.02, prob_loss:   1.82, total_loss:   3.37\n",
            " Total Loss  tf.Tensor(4.4455414, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   24/125, lr:0.000002, giou_loss:   1.90, conf_loss:   0.14, prob_loss:   2.40, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(5.681212, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   25/125, lr:0.000002, giou_loss:   2.70, conf_loss:   0.04, prob_loss:   2.94, total_loss:   5.68\n",
            " Total Loss  tf.Tensor(2.636447, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   26/125, lr:0.000002, giou_loss:   1.24, conf_loss:   0.01, prob_loss:   1.39, total_loss:   2.64\n",
            " Total Loss  tf.Tensor(6.038143, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   27/125, lr:0.000002, giou_loss:   2.77, conf_loss:   0.18, prob_loss:   3.10, total_loss:   6.04\n",
            " Total Loss  tf.Tensor(4.820486, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   28/125, lr:0.000002, giou_loss:   2.27, conf_loss:   0.02, prob_loss:   2.52, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(3.9334412, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   29/125, lr:0.000002, giou_loss:   1.84, conf_loss:   0.11, prob_loss:   1.98, total_loss:   3.93\n",
            " Total Loss  tf.Tensor(3.2469468, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   30/125, lr:0.000002, giou_loss:   1.48, conf_loss:   0.09, prob_loss:   1.68, total_loss:   3.25\n",
            " Total Loss  tf.Tensor(4.6570125, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   31/125, lr:0.000002, giou_loss:   2.04, conf_loss:   0.23, prob_loss:   2.39, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.801546, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   32/125, lr:0.000002, giou_loss:   2.20, conf_loss:   0.17, prob_loss:   2.43, total_loss:   4.80\n",
            " Total Loss  tf.Tensor(4.9396772, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   33/125, lr:0.000002, giou_loss:   2.28, conf_loss:   0.32, prob_loss:   2.34, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(3.8659816, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   34/125, lr:0.000002, giou_loss:   1.64, conf_loss:   0.05, prob_loss:   2.18, total_loss:   3.87\n",
            " Total Loss  tf.Tensor(5.4659634, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   35/125, lr:0.000002, giou_loss:   2.63, conf_loss:   0.03, prob_loss:   2.81, total_loss:   5.47\n",
            " Total Loss  tf.Tensor(4.303053, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   36/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.04, prob_loss:   2.28, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.0562954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   37/125, lr:0.000002, giou_loss:   1.90, conf_loss:   0.10, prob_loss:   2.05, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(4.811942, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   38/125, lr:0.000002, giou_loss:   2.30, conf_loss:   0.33, prob_loss:   2.19, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(3.7522876, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   39/125, lr:0.000002, giou_loss:   1.73, conf_loss:   0.01, prob_loss:   2.01, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(5.3643494, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   40/125, lr:0.000002, giou_loss:   2.53, conf_loss:   0.08, prob_loss:   2.75, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(6.0664296, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   41/125, lr:0.000002, giou_loss:   2.94, conf_loss:   0.38, prob_loss:   2.74, total_loss:   6.07\n",
            " Total Loss  tf.Tensor(4.148787, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   42/125, lr:0.000002, giou_loss:   2.01, conf_loss:   0.04, prob_loss:   2.10, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(4.2805257, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   43/125, lr:0.000002, giou_loss:   1.97, conf_loss:   0.18, prob_loss:   2.12, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(4.684143, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   44/125, lr:0.000002, giou_loss:   1.91, conf_loss:   0.14, prob_loss:   2.64, total_loss:   4.68\n",
            " Total Loss  tf.Tensor(4.8922863, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   45/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.08, prob_loss:   2.74, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(4.791872, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   46/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.07, prob_loss:   2.56, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.770171, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   47/125, lr:0.000002, giou_loss:   2.24, conf_loss:   0.04, prob_loss:   2.48, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(3.229991, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   48/125, lr:0.000002, giou_loss:   1.44, conf_loss:   0.06, prob_loss:   1.73, total_loss:   3.23\n",
            " Total Loss  tf.Tensor(5.066334, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   49/125, lr:0.000002, giou_loss:   2.36, conf_loss:   0.15, prob_loss:   2.56, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(3.7320395, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   50/125, lr:0.000002, giou_loss:   1.68, conf_loss:   0.17, prob_loss:   1.88, total_loss:   3.73\n",
            " Total Loss  tf.Tensor(4.0141015, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   51/125, lr:0.000002, giou_loss:   1.87, conf_loss:   0.01, prob_loss:   2.13, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(4.3572187, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   52/125, lr:0.000002, giou_loss:   2.12, conf_loss:   0.03, prob_loss:   2.21, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(4.7288094, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   53/125, lr:0.000002, giou_loss:   2.21, conf_loss:   0.13, prob_loss:   2.39, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(6.0094786, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   54/125, lr:0.000002, giou_loss:   2.86, conf_loss:   0.21, prob_loss:   2.94, total_loss:   6.01\n",
            " Total Loss  tf.Tensor(4.3466463, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   55/125, lr:0.000002, giou_loss:   2.29, conf_loss:   0.08, prob_loss:   1.98, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(4.288868, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   56/125, lr:0.000002, giou_loss:   1.88, conf_loss:   0.17, prob_loss:   2.24, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.3113756, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   57/125, lr:0.000002, giou_loss:   1.90, conf_loss:   0.01, prob_loss:   2.39, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.1369944, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   58/125, lr:0.000002, giou_loss:   2.00, conf_loss:   0.04, prob_loss:   2.09, total_loss:   4.14\n",
            " Total Loss  tf.Tensor(3.874465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   59/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.36, prob_loss:   1.80, total_loss:   3.87\n",
            " Total Loss  tf.Tensor(5.4380236, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   60/125, lr:0.000002, giou_loss:   2.53, conf_loss:   0.11, prob_loss:   2.79, total_loss:   5.44\n",
            " Total Loss  tf.Tensor(4.754834, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   61/125, lr:0.000002, giou_loss:   2.23, conf_loss:   0.04, prob_loss:   2.48, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(4.1697707, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   62/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.05, prob_loss:   2.05, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(3.865311, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   63/125, lr:0.000002, giou_loss:   1.93, conf_loss:   0.10, prob_loss:   1.83, total_loss:   3.87\n",
            " Total Loss  tf.Tensor(4.6252494, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   64/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.15, prob_loss:   2.32, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(5.5347066, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   65/125, lr:0.000002, giou_loss:   2.53, conf_loss:   0.33, prob_loss:   2.68, total_loss:   5.53\n",
            " Total Loss  tf.Tensor(4.354163, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   66/125, lr:0.000002, giou_loss:   2.03, conf_loss:   0.04, prob_loss:   2.29, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(5.0381374, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   67/125, lr:0.000002, giou_loss:   2.30, conf_loss:   0.13, prob_loss:   2.61, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(4.319483, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   68/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.02, prob_loss:   2.29, total_loss:   4.32\n",
            " Total Loss  tf.Tensor(3.7402792, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   69/125, lr:0.000002, giou_loss:   1.74, conf_loss:   0.12, prob_loss:   1.87, total_loss:   3.74\n",
            " Total Loss  tf.Tensor(4.156681, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   70/125, lr:0.000002, giou_loss:   1.83, conf_loss:   0.17, prob_loss:   2.16, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(4.148718, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   71/125, lr:0.000002, giou_loss:   1.89, conf_loss:   0.07, prob_loss:   2.19, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(4.4830155, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   72/125, lr:0.000002, giou_loss:   2.05, conf_loss:   0.15, prob_loss:   2.28, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(4.3252344, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   73/125, lr:0.000002, giou_loss:   1.86, conf_loss:   0.04, prob_loss:   2.42, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(2.6482697, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   74/125, lr:0.000002, giou_loss:   1.09, conf_loss:   0.01, prob_loss:   1.54, total_loss:   2.65\n",
            " Total Loss  tf.Tensor(4.1766562, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   75/125, lr:0.000002, giou_loss:   1.88, conf_loss:   0.09, prob_loss:   2.21, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(4.0112023, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   76/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.02, prob_loss:   2.20, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(3.2131376, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   77/125, lr:0.000002, giou_loss:   1.35, conf_loss:   0.04, prob_loss:   1.82, total_loss:   3.21\n",
            " Total Loss  tf.Tensor(5.1244307, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   78/125, lr:0.000002, giou_loss:   2.38, conf_loss:   0.08, prob_loss:   2.67, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(5.5030317, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   79/125, lr:0.000002, giou_loss:   2.31, conf_loss:   0.19, prob_loss:   3.01, total_loss:   5.50\n",
            " Total Loss  tf.Tensor(4.212738, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   80/125, lr:0.000002, giou_loss:   1.92, conf_loss:   0.17, prob_loss:   2.12, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(5.4152803, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   81/125, lr:0.000002, giou_loss:   2.46, conf_loss:   0.38, prob_loss:   2.58, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(3.6911306, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   82/125, lr:0.000002, giou_loss:   1.53, conf_loss:   0.04, prob_loss:   2.11, total_loss:   3.69\n",
            " Total Loss  tf.Tensor(5.9065695, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   83/125, lr:0.000002, giou_loss:   2.70, conf_loss:   0.10, prob_loss:   3.11, total_loss:   5.91\n",
            " Total Loss  tf.Tensor(3.8743079, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   84/125, lr:0.000002, giou_loss:   1.66, conf_loss:   0.01, prob_loss:   2.21, total_loss:   3.87\n",
            " Total Loss  tf.Tensor(4.4032817, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   85/125, lr:0.000002, giou_loss:   2.13, conf_loss:   0.04, prob_loss:   2.24, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(3.5994704, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   86/125, lr:0.000002, giou_loss:   1.65, conf_loss:   0.02, prob_loss:   1.93, total_loss:   3.60\n",
            " Total Loss  tf.Tensor(4.286395, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   87/125, lr:0.000002, giou_loss:   1.98, conf_loss:   0.05, prob_loss:   2.26, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(5.9340897, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   88/125, lr:0.000002, giou_loss:   2.70, conf_loss:   0.24, prob_loss:   3.00, total_loss:   5.93\n",
            " Total Loss  tf.Tensor(4.5363684, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   89/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.07, prob_loss:   2.44, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(5.5892105, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   90/125, lr:0.000002, giou_loss:   2.64, conf_loss:   0.06, prob_loss:   2.89, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(3.8446813, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   91/125, lr:0.000002, giou_loss:   1.85, conf_loss:   0.11, prob_loss:   1.88, total_loss:   3.84\n",
            " Total Loss  tf.Tensor(4.5618286, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   92/125, lr:0.000002, giou_loss:   2.14, conf_loss:   0.23, prob_loss:   2.19, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.563963, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   93/125, lr:0.000002, giou_loss:   1.92, conf_loss:   0.06, prob_loss:   2.58, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(5.6634274, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   94/125, lr:0.000002, giou_loss:   2.62, conf_loss:   0.20, prob_loss:   2.85, total_loss:   5.66\n",
            " Total Loss  tf.Tensor(4.6573677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   95/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.08, prob_loss:   2.42, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(3.5340352, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   96/125, lr:0.000002, giou_loss:   1.63, conf_loss:   0.12, prob_loss:   1.78, total_loss:   3.53\n",
            " Total Loss  tf.Tensor(5.3424664, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   97/125, lr:0.000002, giou_loss:   2.47, conf_loss:   0.10, prob_loss:   2.78, total_loss:   5.34\n",
            " Total Loss  tf.Tensor(4.9906564, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   98/125, lr:0.000002, giou_loss:   2.45, conf_loss:   0.17, prob_loss:   2.37, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(4.6483517, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:   99/125, lr:0.000002, giou_loss:   2.01, conf_loss:   0.46, prob_loss:   2.17, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(6.2061377, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  100/125, lr:0.000002, giou_loss:   2.86, conf_loss:   0.21, prob_loss:   3.14, total_loss:   6.21\n",
            " Total Loss  tf.Tensor(4.596299, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  101/125, lr:0.000002, giou_loss:   2.09, conf_loss:   0.08, prob_loss:   2.43, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(4.478113, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  102/125, lr:0.000002, giou_loss:   2.15, conf_loss:   0.05, prob_loss:   2.28, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(5.425958, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  103/125, lr:0.000002, giou_loss:   2.43, conf_loss:   0.03, prob_loss:   2.97, total_loss:   5.43\n",
            " Total Loss  tf.Tensor(5.024856, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  104/125, lr:0.000002, giou_loss:   2.14, conf_loss:   0.20, prob_loss:   2.69, total_loss:   5.02\n",
            " Total Loss  tf.Tensor(4.6434855, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  105/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.05, prob_loss:   2.53, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(5.122175, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  106/125, lr:0.000002, giou_loss:   2.39, conf_loss:   0.06, prob_loss:   2.67, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(3.1354814, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  107/125, lr:0.000002, giou_loss:   1.43, conf_loss:   0.02, prob_loss:   1.69, total_loss:   3.14\n",
            " Total Loss  tf.Tensor(4.7111974, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  108/125, lr:0.000002, giou_loss:   2.00, conf_loss:   0.06, prob_loss:   2.65, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(3.958694, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  109/125, lr:0.000002, giou_loss:   1.86, conf_loss:   0.02, prob_loss:   2.09, total_loss:   3.96\n",
            " Total Loss  tf.Tensor(5.854743, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  110/125, lr:0.000002, giou_loss:   2.70, conf_loss:   0.26, prob_loss:   2.89, total_loss:   5.85\n",
            " Total Loss  tf.Tensor(4.988159, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  111/125, lr:0.000002, giou_loss:   2.34, conf_loss:   0.06, prob_loss:   2.59, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(4.4491754, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  112/125, lr:0.000002, giou_loss:   2.02, conf_loss:   0.23, prob_loss:   2.19, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.509484, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  113/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.10, prob_loss:   2.24, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(5.9188232, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  114/125, lr:0.000002, giou_loss:   2.88, conf_loss:   0.23, prob_loss:   2.80, total_loss:   5.92\n",
            " Total Loss  tf.Tensor(5.2271624, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  115/125, lr:0.000002, giou_loss:   2.37, conf_loss:   0.20, prob_loss:   2.66, total_loss:   5.23\n",
            " Total Loss  tf.Tensor(4.149127, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  116/125, lr:0.000002, giou_loss:   1.78, conf_loss:   0.09, prob_loss:   2.28, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(4.0444613, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  117/125, lr:0.000002, giou_loss:   1.82, conf_loss:   0.07, prob_loss:   2.16, total_loss:   4.04\n",
            " Total Loss  tf.Tensor(5.1781187, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  118/125, lr:0.000002, giou_loss:   2.26, conf_loss:   0.26, prob_loss:   2.66, total_loss:   5.18\n",
            " Total Loss  tf.Tensor(5.497706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  119/125, lr:0.000002, giou_loss:   2.49, conf_loss:   0.21, prob_loss:   2.80, total_loss:   5.50\n",
            " Total Loss  tf.Tensor(3.9684, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  120/125, lr:0.000002, giou_loss:   1.84, conf_loss:   0.05, prob_loss:   2.07, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(3.3934631, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  121/125, lr:0.000002, giou_loss:   1.54, conf_loss:   0.03, prob_loss:   1.82, total_loss:   3.39\n",
            " Total Loss  tf.Tensor(3.9774714, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  122/125, lr:0.000002, giou_loss:   1.68, conf_loss:   0.09, prob_loss:   2.21, total_loss:   3.98\n",
            " Total Loss  tf.Tensor(3.941317, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  123/125, lr:0.000002, giou_loss:   1.78, conf_loss:   0.03, prob_loss:   2.13, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(6.511177, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:  124/125, lr:0.000002, giou_loss:   3.08, conf_loss:   0.31, prob_loss:   3.12, total_loss:   6.51\n",
            " Total Loss  tf.Tensor(5.2243624, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    0/125, lr:0.000002, giou_loss:   2.16, conf_loss:   0.03, prob_loss:   3.04, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(3.6974633, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:94 step:    1/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.07, prob_loss:   1.91, total_loss:   3.70\n",
            "\n",
            "\n",
            "giou_val_loss:   2.19, conf_val_loss:   0.64, prob_val_loss:   5.27, total_val_loss:   8.11\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(5.693448, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    2/125, lr:0.000002, giou_loss:   2.76, conf_loss:   0.06, prob_loss:   2.88, total_loss:   5.69\n",
            " Total Loss  tf.Tensor(4.635957, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    3/125, lr:0.000002, giou_loss:   1.90, conf_loss:   0.44, prob_loss:   2.30, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.3569117, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    4/125, lr:0.000002, giou_loss:   2.17, conf_loss:   0.18, prob_loss:   2.01, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(3.6157532, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    5/125, lr:0.000002, giou_loss:   1.59, conf_loss:   0.05, prob_loss:   1.98, total_loss:   3.62\n",
            " Total Loss  tf.Tensor(5.0550795, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    6/125, lr:0.000002, giou_loss:   2.21, conf_loss:   0.10, prob_loss:   2.74, total_loss:   5.06\n",
            " Total Loss  tf.Tensor(4.929981, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    7/125, lr:0.000002, giou_loss:   2.47, conf_loss:   0.07, prob_loss:   2.39, total_loss:   4.93\n",
            " Total Loss  tf.Tensor(4.7322903, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    8/125, lr:0.000002, giou_loss:   2.17, conf_loss:   0.09, prob_loss:   2.48, total_loss:   4.73\n",
            " Total Loss  tf.Tensor(4.6772566, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    9/125, lr:0.000002, giou_loss:   2.10, conf_loss:   0.03, prob_loss:   2.54, total_loss:   4.68\n",
            " Total Loss  tf.Tensor(4.576729, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   10/125, lr:0.000002, giou_loss:   1.71, conf_loss:   0.07, prob_loss:   2.80, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.655602, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   11/125, lr:0.000002, giou_loss:   2.03, conf_loss:   0.06, prob_loss:   2.56, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(3.9810174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   12/125, lr:0.000002, giou_loss:   1.71, conf_loss:   0.25, prob_loss:   2.02, total_loss:   3.98\n",
            " Total Loss  tf.Tensor(4.5174465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   13/125, lr:0.000002, giou_loss:   2.06, conf_loss:   0.04, prob_loss:   2.41, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(4.804698, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   14/125, lr:0.000002, giou_loss:   2.35, conf_loss:   0.12, prob_loss:   2.33, total_loss:   4.80\n",
            " Total Loss  tf.Tensor(4.623514, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   15/125, lr:0.000002, giou_loss:   2.12, conf_loss:   0.14, prob_loss:   2.36, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(3.8590817, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   16/125, lr:0.000002, giou_loss:   1.51, conf_loss:   0.43, prob_loss:   1.92, total_loss:   3.86\n",
            " Total Loss  tf.Tensor(4.988632, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   17/125, lr:0.000002, giou_loss:   2.37, conf_loss:   0.19, prob_loss:   2.43, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(4.0911016, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   18/125, lr:0.000002, giou_loss:   1.93, conf_loss:   0.08, prob_loss:   2.08, total_loss:   4.09\n",
            " Total Loss  tf.Tensor(4.3003287, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   19/125, lr:0.000002, giou_loss:   2.10, conf_loss:   0.02, prob_loss:   2.18, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.946806, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   20/125, lr:0.000002, giou_loss:   2.41, conf_loss:   0.04, prob_loss:   2.50, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(5.360567, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   21/125, lr:0.000002, giou_loss:   2.44, conf_loss:   0.05, prob_loss:   2.88, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(4.0279417, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   22/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.11, prob_loss:   1.93, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(4.2697773, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   23/125, lr:0.000002, giou_loss:   1.81, conf_loss:   0.02, prob_loss:   2.44, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(6.35679, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   24/125, lr:0.000002, giou_loss:   2.86, conf_loss:   0.46, prob_loss:   3.04, total_loss:   6.36\n",
            " Total Loss  tf.Tensor(3.4143758, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   25/125, lr:0.000002, giou_loss:   1.62, conf_loss:   0.01, prob_loss:   1.79, total_loss:   3.41\n",
            " Total Loss  tf.Tensor(3.4155502, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   26/125, lr:0.000002, giou_loss:   1.76, conf_loss:   0.02, prob_loss:   1.63, total_loss:   3.42\n",
            " Total Loss  tf.Tensor(4.0084286, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   27/125, lr:0.000002, giou_loss:   1.68, conf_loss:   0.32, prob_loss:   2.01, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(4.9519215, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   28/125, lr:0.000002, giou_loss:   2.29, conf_loss:   0.10, prob_loss:   2.56, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(4.221091, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   29/125, lr:0.000002, giou_loss:   1.77, conf_loss:   0.06, prob_loss:   2.40, total_loss:   4.22\n",
            " Total Loss  tf.Tensor(5.1987643, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   30/125, lr:0.000002, giou_loss:   2.42, conf_loss:   0.10, prob_loss:   2.68, total_loss:   5.20\n",
            " Total Loss  tf.Tensor(5.4915533, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   31/125, lr:0.000002, giou_loss:   2.63, conf_loss:   0.31, prob_loss:   2.55, total_loss:   5.49\n",
            " Total Loss  tf.Tensor(4.5525527, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   32/125, lr:0.000002, giou_loss:   2.00, conf_loss:   0.30, prob_loss:   2.25, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(4.064953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   33/125, lr:0.000002, giou_loss:   1.79, conf_loss:   0.16, prob_loss:   2.12, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(5.0783625, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   34/125, lr:0.000002, giou_loss:   2.51, conf_loss:   0.13, prob_loss:   2.44, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(3.3712006, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   35/125, lr:0.000002, giou_loss:   1.44, conf_loss:   0.01, prob_loss:   1.91, total_loss:   3.37\n",
            " Total Loss  tf.Tensor(4.166008, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   36/125, lr:0.000002, giou_loss:   2.09, conf_loss:   0.18, prob_loss:   1.90, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.5872517, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   37/125, lr:0.000002, giou_loss:   1.93, conf_loss:   0.13, prob_loss:   2.53, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(5.257318, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   38/125, lr:0.000002, giou_loss:   2.49, conf_loss:   0.15, prob_loss:   2.62, total_loss:   5.26\n",
            " Total Loss  tf.Tensor(5.364465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   39/125, lr:0.000002, giou_loss:   2.56, conf_loss:   0.07, prob_loss:   2.73, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(4.516695, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   40/125, lr:0.000002, giou_loss:   2.11, conf_loss:   0.05, prob_loss:   2.36, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(6.157323, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   41/125, lr:0.000002, giou_loss:   2.97, conf_loss:   0.15, prob_loss:   3.04, total_loss:   6.16\n",
            " Total Loss  tf.Tensor(4.167283, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   42/125, lr:0.000002, giou_loss:   1.98, conf_loss:   0.04, prob_loss:   2.15, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.154231, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   43/125, lr:0.000002, giou_loss:   1.91, conf_loss:   0.17, prob_loss:   2.07, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(4.4906626, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   44/125, lr:0.000002, giou_loss:   2.22, conf_loss:   0.04, prob_loss:   2.24, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(3.5605865, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   45/125, lr:0.000002, giou_loss:   1.85, conf_loss:   0.06, prob_loss:   1.65, total_loss:   3.56\n",
            " Total Loss  tf.Tensor(3.688685, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   46/125, lr:0.000002, giou_loss:   1.75, conf_loss:   0.10, prob_loss:   1.84, total_loss:   3.69\n",
            " Total Loss  tf.Tensor(4.8205748, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   47/125, lr:0.000002, giou_loss:   2.05, conf_loss:   0.12, prob_loss:   2.65, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(6.026368, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   48/125, lr:0.000002, giou_loss:   2.90, conf_loss:   0.27, prob_loss:   2.85, total_loss:   6.03\n",
            " Total Loss  tf.Tensor(5.9936514, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   49/125, lr:0.000002, giou_loss:   2.65, conf_loss:   0.20, prob_loss:   3.15, total_loss:   5.99\n",
            " Total Loss  tf.Tensor(3.3582466, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   50/125, lr:0.000002, giou_loss:   1.66, conf_loss:   0.07, prob_loss:   1.63, total_loss:   3.36\n",
            " Total Loss  tf.Tensor(4.9514313, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   51/125, lr:0.000002, giou_loss:   2.25, conf_loss:   0.14, prob_loss:   2.56, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(5.2184086, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   52/125, lr:0.000002, giou_loss:   2.29, conf_loss:   0.22, prob_loss:   2.71, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(4.1280546, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   53/125, lr:0.000002, giou_loss:   1.77, conf_loss:   0.01, prob_loss:   2.34, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(5.414954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   54/125, lr:0.000002, giou_loss:   2.64, conf_loss:   0.04, prob_loss:   2.74, total_loss:   5.41\n",
            " Total Loss  tf.Tensor(4.7125883, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   55/125, lr:0.000002, giou_loss:   2.28, conf_loss:   0.22, prob_loss:   2.21, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(3.2842028, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   56/125, lr:0.000002, giou_loss:   1.31, conf_loss:   0.04, prob_loss:   1.93, total_loss:   3.28\n",
            " Total Loss  tf.Tensor(3.5544176, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   57/125, lr:0.000002, giou_loss:   1.72, conf_loss:   0.03, prob_loss:   1.81, total_loss:   3.55\n",
            " Total Loss  tf.Tensor(4.5669856, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   58/125, lr:0.000002, giou_loss:   2.06, conf_loss:   0.05, prob_loss:   2.46, total_loss:   4.57\n",
            " Total Loss  tf.Tensor(3.787178, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   59/125, lr:0.000002, giou_loss:   1.62, conf_loss:   0.22, prob_loss:   1.94, total_loss:   3.79\n",
            " Total Loss  tf.Tensor(3.3091626, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   60/125, lr:0.000002, giou_loss:   1.52, conf_loss:   0.08, prob_loss:   1.71, total_loss:   3.31\n",
            " Total Loss  tf.Tensor(4.7010283, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   61/125, lr:0.000002, giou_loss:   2.19, conf_loss:   0.19, prob_loss:   2.32, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(5.440716, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   62/125, lr:0.000002, giou_loss:   2.53, conf_loss:   0.10, prob_loss:   2.81, total_loss:   5.44\n",
            " Total Loss  tf.Tensor(4.943763, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   63/125, lr:0.000002, giou_loss:   2.29, conf_loss:   0.15, prob_loss:   2.51, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(4.1948957, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   64/125, lr:0.000002, giou_loss:   1.91, conf_loss:   0.06, prob_loss:   2.23, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(4.007082, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   65/125, lr:0.000002, giou_loss:   1.78, conf_loss:   0.02, prob_loss:   2.21, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(4.2250953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   66/125, lr:0.000002, giou_loss:   1.99, conf_loss:   0.07, prob_loss:   2.17, total_loss:   4.23\n",
            " Total Loss  tf.Tensor(4.4565735, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   67/125, lr:0.000002, giou_loss:   2.08, conf_loss:   0.07, prob_loss:   2.31, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(3.3715525, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   68/125, lr:0.000002, giou_loss:   1.64, conf_loss:   0.01, prob_loss:   1.72, total_loss:   3.37\n",
            " Total Loss  tf.Tensor(4.451437, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   69/125, lr:0.000002, giou_loss:   2.07, conf_loss:   0.02, prob_loss:   2.37, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(5.0708237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   70/125, lr:0.000002, giou_loss:   2.21, conf_loss:   0.10, prob_loss:   2.77, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(5.3559294, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   71/125, lr:0.000001, giou_loss:   2.48, conf_loss:   0.18, prob_loss:   2.70, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(4.630549, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   72/125, lr:0.000001, giou_loss:   2.25, conf_loss:   0.03, prob_loss:   2.35, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(3.7710264, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   73/125, lr:0.000001, giou_loss:   1.70, conf_loss:   0.05, prob_loss:   2.02, total_loss:   3.77\n",
            " Total Loss  tf.Tensor(4.207431, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   74/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.28, prob_loss:   1.94, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.342862, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   75/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.11, prob_loss:   2.22, total_loss:   4.34\n",
            " Total Loss  tf.Tensor(3.7518387, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   76/125, lr:0.000001, giou_loss:   1.71, conf_loss:   0.12, prob_loss:   1.92, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(4.110615, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   77/125, lr:0.000001, giou_loss:   1.79, conf_loss:   0.04, prob_loss:   2.29, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(4.0700283, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   78/125, lr:0.000001, giou_loss:   1.91, conf_loss:   0.07, prob_loss:   2.08, total_loss:   4.07\n",
            " Total Loss  tf.Tensor(4.0230193, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   79/125, lr:0.000001, giou_loss:   1.79, conf_loss:   0.03, prob_loss:   2.21, total_loss:   4.02\n",
            " Total Loss  tf.Tensor(5.390671, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   80/125, lr:0.000001, giou_loss:   2.41, conf_loss:   0.09, prob_loss:   2.88, total_loss:   5.39\n",
            " Total Loss  tf.Tensor(6.327885, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   81/125, lr:0.000001, giou_loss:   2.41, conf_loss:   0.25, prob_loss:   3.67, total_loss:   6.33\n",
            " Total Loss  tf.Tensor(4.2683268, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   82/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.05, prob_loss:   2.22, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(5.0482483, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   83/125, lr:0.000001, giou_loss:   2.16, conf_loss:   0.19, prob_loss:   2.70, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(4.1036305, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   84/125, lr:0.000001, giou_loss:   1.90, conf_loss:   0.05, prob_loss:   2.16, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(4.71426, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   85/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.14, prob_loss:   2.52, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(4.0484533, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   86/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.04, prob_loss:   2.18, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(5.1764555, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   87/125, lr:0.000001, giou_loss:   1.44, conf_loss:   0.01, prob_loss:   3.72, total_loss:   5.18\n",
            " Total Loss  tf.Tensor(4.660481, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   88/125, lr:0.000001, giou_loss:   2.16, conf_loss:   0.05, prob_loss:   2.45, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.637907, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   89/125, lr:0.000001, giou_loss:   2.15, conf_loss:   0.18, prob_loss:   2.31, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.4090643, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   90/125, lr:0.000001, giou_loss:   1.95, conf_loss:   0.08, prob_loss:   2.38, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(3.534436, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   91/125, lr:0.000001, giou_loss:   1.90, conf_loss:   0.05, prob_loss:   1.59, total_loss:   3.53\n",
            " Total Loss  tf.Tensor(5.355514, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   92/125, lr:0.000001, giou_loss:   2.40, conf_loss:   0.29, prob_loss:   2.67, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(3.7662141, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   93/125, lr:0.000001, giou_loss:   1.63, conf_loss:   0.13, prob_loss:   2.01, total_loss:   3.77\n",
            " Total Loss  tf.Tensor(5.61232, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   94/125, lr:0.000001, giou_loss:   2.70, conf_loss:   0.08, prob_loss:   2.83, total_loss:   5.61\n",
            " Total Loss  tf.Tensor(5.2107906, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   95/125, lr:0.000001, giou_loss:   2.51, conf_loss:   0.08, prob_loss:   2.62, total_loss:   5.21\n",
            " Total Loss  tf.Tensor(4.5792155, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   96/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.09, prob_loss:   2.50, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.5663142, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   97/125, lr:0.000001, giou_loss:   1.67, conf_loss:   0.10, prob_loss:   1.80, total_loss:   3.57\n",
            " Total Loss  tf.Tensor(5.819499, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   98/125, lr:0.000001, giou_loss:   2.73, conf_loss:   0.03, prob_loss:   3.06, total_loss:   5.82\n",
            " Total Loss  tf.Tensor(4.14718, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:   99/125, lr:0.000001, giou_loss:   1.84, conf_loss:   0.08, prob_loss:   2.23, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(3.9844346, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  100/125, lr:0.000001, giou_loss:   1.67, conf_loss:   0.05, prob_loss:   2.26, total_loss:   3.98\n",
            " Total Loss  tf.Tensor(4.379001, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  101/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.12, prob_loss:   2.24, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(4.5453405, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  102/125, lr:0.000001, giou_loss:   1.96, conf_loss:   0.24, prob_loss:   2.35, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(4.634428, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  103/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.17, prob_loss:   2.40, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(3.2836452, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  104/125, lr:0.000001, giou_loss:   1.40, conf_loss:   0.02, prob_loss:   1.86, total_loss:   3.28\n",
            " Total Loss  tf.Tensor(5.7201834, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  105/125, lr:0.000001, giou_loss:   2.67, conf_loss:   0.20, prob_loss:   2.86, total_loss:   5.72\n",
            " Total Loss  tf.Tensor(4.6068897, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  106/125, lr:0.000001, giou_loss:   2.11, conf_loss:   0.16, prob_loss:   2.34, total_loss:   4.61\n",
            " Total Loss  tf.Tensor(5.100846, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  107/125, lr:0.000001, giou_loss:   2.37, conf_loss:   0.05, prob_loss:   2.68, total_loss:   5.10\n",
            " Total Loss  tf.Tensor(5.0792212, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  108/125, lr:0.000001, giou_loss:   2.26, conf_loss:   0.23, prob_loss:   2.59, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(5.14991, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  109/125, lr:0.000001, giou_loss:   2.34, conf_loss:   0.03, prob_loss:   2.78, total_loss:   5.15\n",
            " Total Loss  tf.Tensor(4.4898458, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  110/125, lr:0.000001, giou_loss:   1.92, conf_loss:   0.04, prob_loss:   2.53, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(3.9495723, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  111/125, lr:0.000001, giou_loss:   1.70, conf_loss:   0.04, prob_loss:   2.22, total_loss:   3.95\n",
            " Total Loss  tf.Tensor(4.2930717, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  112/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.01, prob_loss:   2.31, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(3.687333, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  113/125, lr:0.000001, giou_loss:   1.57, conf_loss:   0.10, prob_loss:   2.02, total_loss:   3.69\n",
            " Total Loss  tf.Tensor(4.87646, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  114/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.24, prob_loss:   2.49, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(3.7878904, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  115/125, lr:0.000001, giou_loss:   1.76, conf_loss:   0.06, prob_loss:   1.96, total_loss:   3.79\n",
            " Total Loss  tf.Tensor(4.1831636, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  116/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.18, prob_loss:   2.02, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(3.7406783, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  117/125, lr:0.000001, giou_loss:   1.73, conf_loss:   0.05, prob_loss:   1.96, total_loss:   3.74\n",
            " Total Loss  tf.Tensor(5.4527264, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  118/125, lr:0.000001, giou_loss:   2.60, conf_loss:   0.09, prob_loss:   2.76, total_loss:   5.45\n",
            " Total Loss  tf.Tensor(4.373451, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  119/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.03, prob_loss:   2.29, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(3.579548, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  120/125, lr:0.000001, giou_loss:   1.74, conf_loss:   0.03, prob_loss:   1.82, total_loss:   3.58\n",
            " Total Loss  tf.Tensor(4.247261, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  121/125, lr:0.000001, giou_loss:   2.01, conf_loss:   0.10, prob_loss:   2.13, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(4.8160877, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  122/125, lr:0.000001, giou_loss:   2.20, conf_loss:   0.20, prob_loss:   2.42, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(3.8168788, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  123/125, lr:0.000001, giou_loss:   1.71, conf_loss:   0.04, prob_loss:   2.07, total_loss:   3.82\n",
            " Total Loss  tf.Tensor(6.975177, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:  124/125, lr:0.000001, giou_loss:   3.29, conf_loss:   0.14, prob_loss:   3.55, total_loss:   6.98\n",
            " Total Loss  tf.Tensor(4.83803, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    0/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.04, prob_loss:   2.58, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(5.0216503, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:95 step:    1/125, lr:0.000001, giou_loss:   2.35, conf_loss:   0.04, prob_loss:   2.63, total_loss:   5.02\n",
            "\n",
            "\n",
            "giou_val_loss:   2.19, conf_val_loss:   0.62, prob_val_loss:   5.31, total_val_loss:   8.12\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.3800497, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    2/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.15, prob_loss:   2.30, total_loss:   4.38\n",
            " Total Loss  tf.Tensor(3.993261, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    3/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.01, prob_loss:   2.18, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(5.458922, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    4/125, lr:0.000001, giou_loss:   2.82, conf_loss:   0.09, prob_loss:   2.55, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(4.2104497, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    5/125, lr:0.000001, giou_loss:   1.90, conf_loss:   0.11, prob_loss:   2.20, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(3.7229314, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    6/125, lr:0.000001, giou_loss:   1.60, conf_loss:   0.03, prob_loss:   2.09, total_loss:   3.72\n",
            " Total Loss  tf.Tensor(3.999255, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    7/125, lr:0.000001, giou_loss:   1.85, conf_loss:   0.07, prob_loss:   2.08, total_loss:   4.00\n",
            " Total Loss  tf.Tensor(4.21149, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    8/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.06, prob_loss:   2.18, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(5.2572637, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    9/125, lr:0.000001, giou_loss:   2.30, conf_loss:   0.20, prob_loss:   2.75, total_loss:   5.26\n",
            " Total Loss  tf.Tensor(4.204211, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   10/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.02, prob_loss:   2.15, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(4.7780375, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   11/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.10, prob_loss:   2.44, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(4.0561514, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   12/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.20, prob_loss:   1.86, total_loss:   4.06\n",
            " Total Loss  tf.Tensor(3.8782647, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   13/125, lr:0.000001, giou_loss:   1.56, conf_loss:   0.02, prob_loss:   2.30, total_loss:   3.88\n",
            " Total Loss  tf.Tensor(4.532655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   14/125, lr:0.000001, giou_loss:   1.95, conf_loss:   0.07, prob_loss:   2.51, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(3.9083343, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   15/125, lr:0.000001, giou_loss:   1.83, conf_loss:   0.07, prob_loss:   2.01, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.3708334, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   16/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.17, prob_loss:   2.27, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(4.468325, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   17/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.02, prob_loss:   2.40, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(5.521985, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   18/125, lr:0.000001, giou_loss:   2.59, conf_loss:   0.19, prob_loss:   2.74, total_loss:   5.52\n",
            " Total Loss  tf.Tensor(6.8370037, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   19/125, lr:0.000001, giou_loss:   2.42, conf_loss:   0.10, prob_loss:   4.32, total_loss:   6.84\n",
            " Total Loss  tf.Tensor(3.5773246, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   20/125, lr:0.000001, giou_loss:   1.74, conf_loss:   0.03, prob_loss:   1.81, total_loss:   3.58\n",
            " Total Loss  tf.Tensor(4.1556206, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   21/125, lr:0.000001, giou_loss:   2.01, conf_loss:   0.03, prob_loss:   2.12, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(3.6999063, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   22/125, lr:0.000001, giou_loss:   1.71, conf_loss:   0.03, prob_loss:   1.96, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(3.426163, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   23/125, lr:0.000001, giou_loss:   1.49, conf_loss:   0.01, prob_loss:   1.93, total_loss:   3.43\n",
            " Total Loss  tf.Tensor(5.6273293, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   24/125, lr:0.000001, giou_loss:   2.71, conf_loss:   0.09, prob_loss:   2.82, total_loss:   5.63\n",
            " Total Loss  tf.Tensor(4.2580986, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   25/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.12, prob_loss:   2.09, total_loss:   4.26\n",
            " Total Loss  tf.Tensor(4.9264326, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   26/125, lr:0.000001, giou_loss:   2.25, conf_loss:   0.05, prob_loss:   2.63, total_loss:   4.93\n",
            " Total Loss  tf.Tensor(3.1192222, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   27/125, lr:0.000001, giou_loss:   1.27, conf_loss:   0.02, prob_loss:   1.84, total_loss:   3.12\n",
            " Total Loss  tf.Tensor(4.4828243, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   28/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.13, prob_loss:   2.22, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(5.837043, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   29/125, lr:0.000001, giou_loss:   2.83, conf_loss:   0.51, prob_loss:   2.50, total_loss:   5.84\n",
            " Total Loss  tf.Tensor(4.809973, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   30/125, lr:0.000001, giou_loss:   2.15, conf_loss:   0.19, prob_loss:   2.47, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(4.4658685, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   31/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.16, prob_loss:   2.20, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(3.78125, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   32/125, lr:0.000001, giou_loss:   1.56, conf_loss:   0.04, prob_loss:   2.19, total_loss:   3.78\n",
            " Total Loss  tf.Tensor(4.788454, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   33/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.10, prob_loss:   2.56, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(3.4756808, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   34/125, lr:0.000001, giou_loss:   1.62, conf_loss:   0.04, prob_loss:   1.82, total_loss:   3.48\n",
            " Total Loss  tf.Tensor(4.8396215, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   35/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.03, prob_loss:   2.49, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.956251, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   36/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.11, prob_loss:   2.53, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(4.8422527, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   37/125, lr:0.000001, giou_loss:   2.48, conf_loss:   0.15, prob_loss:   2.21, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(5.303293, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   38/125, lr:0.000001, giou_loss:   2.60, conf_loss:   0.22, prob_loss:   2.48, total_loss:   5.30\n",
            " Total Loss  tf.Tensor(5.1997366, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   39/125, lr:0.000001, giou_loss:   2.66, conf_loss:   0.14, prob_loss:   2.40, total_loss:   5.20\n",
            " Total Loss  tf.Tensor(4.1174293, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   40/125, lr:0.000001, giou_loss:   1.85, conf_loss:   0.09, prob_loss:   2.17, total_loss:   4.12\n",
            " Total Loss  tf.Tensor(4.205961, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   41/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.07, prob_loss:   2.12, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(5.582565, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   42/125, lr:0.000001, giou_loss:   2.47, conf_loss:   0.04, prob_loss:   3.07, total_loss:   5.58\n",
            " Total Loss  tf.Tensor(4.57644, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   43/125, lr:0.000001, giou_loss:   2.29, conf_loss:   0.12, prob_loss:   2.17, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.522705, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   44/125, lr:0.000001, giou_loss:   1.64, conf_loss:   0.04, prob_loss:   1.84, total_loss:   3.52\n",
            " Total Loss  tf.Tensor(4.5573835, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   45/125, lr:0.000001, giou_loss:   2.11, conf_loss:   0.31, prob_loss:   2.14, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.6362467, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   46/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.04, prob_loss:   2.52, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(3.8103266, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   47/125, lr:0.000001, giou_loss:   1.85, conf_loss:   0.02, prob_loss:   1.94, total_loss:   3.81\n",
            " Total Loss  tf.Tensor(3.9231706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   48/125, lr:0.000001, giou_loss:   1.73, conf_loss:   0.06, prob_loss:   2.13, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(4.5867653, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   49/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.06, prob_loss:   2.45, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(3.84706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   50/125, lr:0.000001, giou_loss:   1.78, conf_loss:   0.03, prob_loss:   2.04, total_loss:   3.85\n",
            " Total Loss  tf.Tensor(3.7290282, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   51/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.05, prob_loss:   1.93, total_loss:   3.73\n",
            " Total Loss  tf.Tensor(5.7786026, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   52/125, lr:0.000001, giou_loss:   2.88, conf_loss:   0.05, prob_loss:   2.84, total_loss:   5.78\n",
            " Total Loss  tf.Tensor(5.071965, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   53/125, lr:0.000001, giou_loss:   2.49, conf_loss:   0.07, prob_loss:   2.52, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(4.6014557, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   54/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.09, prob_loss:   2.52, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(5.4294376, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   55/125, lr:0.000001, giou_loss:   2.55, conf_loss:   0.08, prob_loss:   2.80, total_loss:   5.43\n",
            " Total Loss  tf.Tensor(5.6157103, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   56/125, lr:0.000001, giou_loss:   2.69, conf_loss:   0.15, prob_loss:   2.78, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(5.65893, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   57/125, lr:0.000001, giou_loss:   2.45, conf_loss:   0.23, prob_loss:   2.98, total_loss:   5.66\n",
            " Total Loss  tf.Tensor(5.5897017, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   58/125, lr:0.000001, giou_loss:   2.82, conf_loss:   0.05, prob_loss:   2.72, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(4.5713315, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   59/125, lr:0.000001, giou_loss:   2.23, conf_loss:   0.09, prob_loss:   2.26, total_loss:   4.57\n",
            " Total Loss  tf.Tensor(4.2452497, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   60/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.06, prob_loss:   2.36, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(3.783607, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   61/125, lr:0.000001, giou_loss:   1.62, conf_loss:   0.15, prob_loss:   2.01, total_loss:   3.78\n",
            " Total Loss  tf.Tensor(5.214363, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   62/125, lr:0.000001, giou_loss:   2.36, conf_loss:   0.51, prob_loss:   2.35, total_loss:   5.21\n",
            " Total Loss  tf.Tensor(4.262638, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   63/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.12, prob_loss:   2.18, total_loss:   4.26\n",
            " Total Loss  tf.Tensor(3.5028257, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   64/125, lr:0.000001, giou_loss:   1.63, conf_loss:   0.08, prob_loss:   1.80, total_loss:   3.50\n",
            " Total Loss  tf.Tensor(5.2074556, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   65/125, lr:0.000001, giou_loss:   2.50, conf_loss:   0.04, prob_loss:   2.66, total_loss:   5.21\n",
            " Total Loss  tf.Tensor(3.7476058, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   66/125, lr:0.000001, giou_loss:   1.79, conf_loss:   0.03, prob_loss:   1.93, total_loss:   3.75\n",
            " Total Loss  tf.Tensor(4.505559, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   67/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.29, prob_loss:   2.28, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.4101706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   68/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.14, prob_loss:   2.22, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(4.5689044, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   69/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.15, prob_loss:   2.29, total_loss:   4.57\n",
            " Total Loss  tf.Tensor(6.1326838, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   70/125, lr:0.000001, giou_loss:   2.99, conf_loss:   0.15, prob_loss:   2.99, total_loss:   6.13\n",
            " Total Loss  tf.Tensor(3.2940292, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   71/125, lr:0.000001, giou_loss:   1.58, conf_loss:   0.03, prob_loss:   1.68, total_loss:   3.29\n",
            " Total Loss  tf.Tensor(5.590906, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   72/125, lr:0.000001, giou_loss:   2.66, conf_loss:   0.24, prob_loss:   2.68, total_loss:   5.59\n",
            " Total Loss  tf.Tensor(5.3804207, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   73/125, lr:0.000001, giou_loss:   2.53, conf_loss:   0.08, prob_loss:   2.77, total_loss:   5.38\n",
            " Total Loss  tf.Tensor(3.2384975, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   74/125, lr:0.000001, giou_loss:   1.42, conf_loss:   0.10, prob_loss:   1.72, total_loss:   3.24\n",
            " Total Loss  tf.Tensor(4.838219, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   75/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.05, prob_loss:   2.47, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.7565355, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   76/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.03, prob_loss:   2.52, total_loss:   4.76\n",
            " Total Loss  tf.Tensor(4.569051, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   77/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.05, prob_loss:   2.30, total_loss:   4.57\n",
            " Total Loss  tf.Tensor(4.1285787, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   78/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.03, prob_loss:   2.28, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(4.0013127, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   79/125, lr:0.000001, giou_loss:   1.72, conf_loss:   0.09, prob_loss:   2.19, total_loss:   4.00\n",
            " Total Loss  tf.Tensor(4.3042574, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   80/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.08, prob_loss:   2.23, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.196289, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   81/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.08, prob_loss:   2.19, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(5.4783983, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   82/125, lr:0.000001, giou_loss:   2.48, conf_loss:   0.12, prob_loss:   2.87, total_loss:   5.48\n",
            " Total Loss  tf.Tensor(5.705968, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   83/125, lr:0.000001, giou_loss:   2.70, conf_loss:   0.19, prob_loss:   2.82, total_loss:   5.71\n",
            " Total Loss  tf.Tensor(3.845362, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   84/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.15, prob_loss:   1.94, total_loss:   3.85\n",
            " Total Loss  tf.Tensor(4.1475945, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   85/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.04, prob_loss:   2.16, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(3.840804, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   86/125, lr:0.000001, giou_loss:   1.81, conf_loss:   0.03, prob_loss:   2.00, total_loss:   3.84\n",
            " Total Loss  tf.Tensor(4.5801287, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   87/125, lr:0.000001, giou_loss:   2.19, conf_loss:   0.08, prob_loss:   2.31, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.5845928, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   88/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.13, prob_loss:   1.79, total_loss:   3.58\n",
            " Total Loss  tf.Tensor(5.928798, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   89/125, lr:0.000001, giou_loss:   2.77, conf_loss:   0.07, prob_loss:   3.09, total_loss:   5.93\n",
            " Total Loss  tf.Tensor(5.455803, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   90/125, lr:0.000001, giou_loss:   2.52, conf_loss:   0.04, prob_loss:   2.90, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(5.455491, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   91/125, lr:0.000001, giou_loss:   2.53, conf_loss:   0.18, prob_loss:   2.74, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(4.964465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   92/125, lr:0.000001, giou_loss:   2.15, conf_loss:   0.13, prob_loss:   2.69, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(4.6044173, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   93/125, lr:0.000001, giou_loss:   2.22, conf_loss:   0.06, prob_loss:   2.32, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(3.7447767, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   94/125, lr:0.000001, giou_loss:   1.70, conf_loss:   0.06, prob_loss:   1.99, total_loss:   3.74\n",
            " Total Loss  tf.Tensor(4.804714, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   95/125, lr:0.000001, giou_loss:   2.34, conf_loss:   0.10, prob_loss:   2.36, total_loss:   4.80\n",
            " Total Loss  tf.Tensor(5.5000215, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   96/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.08, prob_loss:   3.28, total_loss:   5.50\n",
            " Total Loss  tf.Tensor(4.018951, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   97/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.02, prob_loss:   2.25, total_loss:   4.02\n",
            " Total Loss  tf.Tensor(3.603374, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   98/125, lr:0.000001, giou_loss:   1.76, conf_loss:   0.04, prob_loss:   1.80, total_loss:   3.60\n",
            " Total Loss  tf.Tensor(4.600543, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:   99/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.42, prob_loss:   2.19, total_loss:   4.60\n",
            " Total Loss  tf.Tensor(4.6804237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  100/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.14, prob_loss:   2.40, total_loss:   4.68\n",
            " Total Loss  tf.Tensor(5.09223, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  101/125, lr:0.000001, giou_loss:   2.37, conf_loss:   0.08, prob_loss:   2.65, total_loss:   5.09\n",
            " Total Loss  tf.Tensor(4.4259787, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  102/125, lr:0.000001, giou_loss:   1.89, conf_loss:   0.09, prob_loss:   2.45, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(5.0435896, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  103/125, lr:0.000001, giou_loss:   2.31, conf_loss:   0.08, prob_loss:   2.66, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(4.898508, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  104/125, lr:0.000001, giou_loss:   2.25, conf_loss:   0.12, prob_loss:   2.53, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(5.1572285, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  105/125, lr:0.000001, giou_loss:   2.39, conf_loss:   0.02, prob_loss:   2.74, total_loss:   5.16\n",
            " Total Loss  tf.Tensor(5.6378055, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  106/125, lr:0.000001, giou_loss:   2.82, conf_loss:   0.04, prob_loss:   2.78, total_loss:   5.64\n",
            " Total Loss  tf.Tensor(3.9001808, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  107/125, lr:0.000001, giou_loss:   1.71, conf_loss:   0.09, prob_loss:   2.10, total_loss:   3.90\n",
            " Total Loss  tf.Tensor(4.6332564, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  108/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.03, prob_loss:   2.51, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(4.2694244, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  109/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.25, prob_loss:   1.91, total_loss:   4.27\n",
            " Total Loss  tf.Tensor(3.1360004, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  110/125, lr:0.000001, giou_loss:   1.56, conf_loss:   0.03, prob_loss:   1.54, total_loss:   3.14\n",
            " Total Loss  tf.Tensor(4.1707516, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  111/125, lr:0.000001, giou_loss:   2.04, conf_loss:   0.09, prob_loss:   2.04, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.9641414, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  112/125, lr:0.000001, giou_loss:   2.41, conf_loss:   0.02, prob_loss:   2.54, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(3.7190576, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  113/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.02, prob_loss:   1.88, total_loss:   3.72\n",
            " Total Loss  tf.Tensor(3.3512821, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  114/125, lr:0.000001, giou_loss:   1.39, conf_loss:   0.12, prob_loss:   1.84, total_loss:   3.35\n",
            " Total Loss  tf.Tensor(4.657148, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  115/125, lr:0.000001, giou_loss:   2.11, conf_loss:   0.10, prob_loss:   2.45, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(3.8887076, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  116/125, lr:0.000001, giou_loss:   1.73, conf_loss:   0.04, prob_loss:   2.11, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(4.6778407, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  117/125, lr:0.000001, giou_loss:   2.18, conf_loss:   0.12, prob_loss:   2.38, total_loss:   4.68\n",
            " Total Loss  tf.Tensor(4.999482, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  118/125, lr:0.000001, giou_loss:   2.33, conf_loss:   0.40, prob_loss:   2.27, total_loss:   5.00\n",
            " Total Loss  tf.Tensor(4.0496025, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  119/125, lr:0.000001, giou_loss:   1.76, conf_loss:   0.01, prob_loss:   2.27, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(4.4045897, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  120/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.03, prob_loss:   2.36, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(4.9230766, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  121/125, lr:0.000001, giou_loss:   2.56, conf_loss:   0.02, prob_loss:   2.35, total_loss:   4.92\n",
            " Total Loss  tf.Tensor(5.0700135, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  122/125, lr:0.000001, giou_loss:   2.31, conf_loss:   0.11, prob_loss:   2.65, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(4.454053, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  123/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.06, prob_loss:   2.25, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.9784327, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:  124/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.09, prob_loss:   2.75, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(4.388856, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    0/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.12, prob_loss:   2.28, total_loss:   4.39\n",
            " Total Loss  tf.Tensor(3.2729063, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:96 step:    1/125, lr:0.000001, giou_loss:   1.51, conf_loss:   0.04, prob_loss:   1.72, total_loss:   3.27\n",
            "\n",
            "\n",
            "giou_val_loss:   2.22, conf_val_loss:   0.65, prob_val_loss:   5.33, total_val_loss:   8.20\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.0874567, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    2/125, lr:0.000001, giou_loss:   1.92, conf_loss:   0.08, prob_loss:   2.10, total_loss:   4.09\n",
            " Total Loss  tf.Tensor(4.1895666, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    3/125, lr:0.000001, giou_loss:   1.89, conf_loss:   0.03, prob_loss:   2.26, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(5.8234797, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    4/125, lr:0.000001, giou_loss:   2.59, conf_loss:   0.28, prob_loss:   2.95, total_loss:   5.82\n",
            " Total Loss  tf.Tensor(4.082553, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    5/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.11, prob_loss:   1.97, total_loss:   4.08\n",
            " Total Loss  tf.Tensor(4.7949333, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    6/125, lr:0.000001, giou_loss:   2.17, conf_loss:   0.19, prob_loss:   2.43, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(3.8172488, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    7/125, lr:0.000001, giou_loss:   1.90, conf_loss:   0.04, prob_loss:   1.87, total_loss:   3.82\n",
            " Total Loss  tf.Tensor(5.248777, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    8/125, lr:0.000001, giou_loss:   2.56, conf_loss:   0.08, prob_loss:   2.61, total_loss:   5.25\n",
            " Total Loss  tf.Tensor(4.286231, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    9/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.10, prob_loss:   2.24, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.7789755, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   10/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.29, prob_loss:   2.35, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(4.8880005, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   11/125, lr:0.000001, giou_loss:   2.19, conf_loss:   0.02, prob_loss:   2.68, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(4.650138, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   12/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.03, prob_loss:   2.57, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(4.752891, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   13/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.14, prob_loss:   2.38, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(3.612188, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   14/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.01, prob_loss:   1.94, total_loss:   3.61\n",
            " Total Loss  tf.Tensor(3.7875118, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   15/125, lr:0.000001, giou_loss:   1.55, conf_loss:   0.27, prob_loss:   1.97, total_loss:   3.79\n",
            " Total Loss  tf.Tensor(3.957907, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   16/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.03, prob_loss:   2.12, total_loss:   3.96\n",
            " Total Loss  tf.Tensor(4.671983, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   17/125, lr:0.000001, giou_loss:   2.27, conf_loss:   0.11, prob_loss:   2.30, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.189955, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   18/125, lr:0.000001, giou_loss:   2.31, conf_loss:   0.03, prob_loss:   2.85, total_loss:   5.19\n",
            " Total Loss  tf.Tensor(4.6716266, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   19/125, lr:0.000001, giou_loss:   2.15, conf_loss:   0.03, prob_loss:   2.49, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.0473437, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   20/125, lr:0.000001, giou_loss:   2.33, conf_loss:   0.69, prob_loss:   2.03, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(4.9525633, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   21/125, lr:0.000001, giou_loss:   2.22, conf_loss:   0.04, prob_loss:   2.69, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(5.558752, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   22/125, lr:0.000001, giou_loss:   2.50, conf_loss:   0.39, prob_loss:   2.67, total_loss:   5.56\n",
            " Total Loss  tf.Tensor(4.2929864, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   23/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.05, prob_loss:   2.22, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.8418655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   24/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.14, prob_loss:   2.58, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.5066233, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   25/125, lr:0.000001, giou_loss:   2.19, conf_loss:   0.08, prob_loss:   2.23, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.306154, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   26/125, lr:0.000001, giou_loss:   1.83, conf_loss:   0.37, prob_loss:   2.11, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.6707296, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   27/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.05, prob_loss:   2.69, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.0366592, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   28/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.11, prob_loss:   2.64, total_loss:   5.04\n",
            " Total Loss  tf.Tensor(4.575285, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   29/125, lr:0.000001, giou_loss:   2.22, conf_loss:   0.04, prob_loss:   2.31, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(5.078574, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   30/125, lr:0.000001, giou_loss:   2.11, conf_loss:   0.29, prob_loss:   2.68, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(4.1794567, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   31/125, lr:0.000001, giou_loss:   1.85, conf_loss:   0.06, prob_loss:   2.27, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(3.2099733, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   32/125, lr:0.000001, giou_loss:   1.45, conf_loss:   0.03, prob_loss:   1.73, total_loss:   3.21\n",
            " Total Loss  tf.Tensor(6.570917, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   33/125, lr:0.000001, giou_loss:   3.01, conf_loss:   0.16, prob_loss:   3.41, total_loss:   6.57\n",
            " Total Loss  tf.Tensor(5.0541677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   34/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.14, prob_loss:   2.60, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(4.0798907, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   35/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.03, prob_loss:   2.04, total_loss:   4.08\n",
            " Total Loss  tf.Tensor(4.6356454, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   36/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.05, prob_loss:   2.35, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(3.9409633, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   37/125, lr:0.000001, giou_loss:   1.87, conf_loss:   0.08, prob_loss:   1.99, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(5.5694466, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   38/125, lr:0.000001, giou_loss:   2.63, conf_loss:   0.08, prob_loss:   2.85, total_loss:   5.57\n",
            " Total Loss  tf.Tensor(3.5243716, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   39/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.02, prob_loss:   1.85, total_loss:   3.52\n",
            " Total Loss  tf.Tensor(4.8813834, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   40/125, lr:0.000001, giou_loss:   2.23, conf_loss:   0.05, prob_loss:   2.60, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(4.186303, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   41/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.09, prob_loss:   2.11, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(6.036352, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   42/125, lr:0.000001, giou_loss:   3.02, conf_loss:   0.23, prob_loss:   2.78, total_loss:   6.04\n",
            " Total Loss  tf.Tensor(3.4647942, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   43/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.02, prob_loss:   1.78, total_loss:   3.46\n",
            " Total Loss  tf.Tensor(4.1897893, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   44/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.07, prob_loss:   2.19, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(4.283572, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   45/125, lr:0.000001, giou_loss:   2.09, conf_loss:   0.06, prob_loss:   2.13, total_loss:   4.28\n",
            " Total Loss  tf.Tensor(4.1031456, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   46/125, lr:0.000001, giou_loss:   1.89, conf_loss:   0.05, prob_loss:   2.16, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(4.405461, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   47/125, lr:0.000001, giou_loss:   2.08, conf_loss:   0.03, prob_loss:   2.29, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(5.297458, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   48/125, lr:0.000001, giou_loss:   2.42, conf_loss:   0.04, prob_loss:   2.83, total_loss:   5.30\n",
            " Total Loss  tf.Tensor(4.6563997, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   49/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.04, prob_loss:   2.55, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.217601, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   50/125, lr:0.000001, giou_loss:   1.68, conf_loss:   0.12, prob_loss:   2.42, total_loss:   4.22\n",
            " Total Loss  tf.Tensor(4.982775, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   51/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.16, prob_loss:   2.50, total_loss:   4.98\n",
            " Total Loss  tf.Tensor(5.8459187, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   52/125, lr:0.000001, giou_loss:   2.62, conf_loss:   0.28, prob_loss:   2.95, total_loss:   5.85\n",
            " Total Loss  tf.Tensor(4.209255, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   53/125, lr:0.000001, giou_loss:   1.95, conf_loss:   0.14, prob_loss:   2.12, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.8242683, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   54/125, lr:0.000001, giou_loss:   2.25, conf_loss:   0.08, prob_loss:   2.50, total_loss:   4.82\n",
            " Total Loss  tf.Tensor(5.336909, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   55/125, lr:0.000001, giou_loss:   2.65, conf_loss:   0.12, prob_loss:   2.57, total_loss:   5.34\n",
            " Total Loss  tf.Tensor(4.992465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   56/125, lr:0.000001, giou_loss:   2.39, conf_loss:   0.12, prob_loss:   2.48, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(4.841961, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   57/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.07, prob_loss:   2.65, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.019665, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   58/125, lr:0.000001, giou_loss:   1.88, conf_loss:   0.05, prob_loss:   2.09, total_loss:   4.02\n",
            " Total Loss  tf.Tensor(3.2775707, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   59/125, lr:0.000001, giou_loss:   1.59, conf_loss:   0.08, prob_loss:   1.61, total_loss:   3.28\n",
            " Total Loss  tf.Tensor(4.8411994, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   60/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.02, prob_loss:   2.50, total_loss:   4.84\n",
            " Total Loss  tf.Tensor(4.947259, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   61/125, lr:0.000001, giou_loss:   2.23, conf_loss:   0.13, prob_loss:   2.58, total_loss:   4.95\n",
            " Total Loss  tf.Tensor(5.220963, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   62/125, lr:0.000001, giou_loss:   2.39, conf_loss:   0.27, prob_loss:   2.56, total_loss:   5.22\n",
            " Total Loss  tf.Tensor(4.6411867, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   63/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.24, prob_loss:   2.35, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.645891, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   64/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.21, prob_loss:   2.42, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(3.8160338, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   65/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.04, prob_loss:   2.02, total_loss:   3.82\n",
            " Total Loss  tf.Tensor(4.484719, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   66/125, lr:0.000001, giou_loss:   2.16, conf_loss:   0.10, prob_loss:   2.23, total_loss:   4.48\n",
            " Total Loss  tf.Tensor(3.2379684, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   67/125, lr:0.000001, giou_loss:   1.43, conf_loss:   0.10, prob_loss:   1.71, total_loss:   3.24\n",
            " Total Loss  tf.Tensor(6.4193325, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   68/125, lr:0.000001, giou_loss:   3.24, conf_loss:   0.19, prob_loss:   2.99, total_loss:   6.42\n",
            " Total Loss  tf.Tensor(4.4334297, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   69/125, lr:0.000001, giou_loss:   2.11, conf_loss:   0.01, prob_loss:   2.32, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(5.1986217, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   70/125, lr:0.000001, giou_loss:   2.60, conf_loss:   0.05, prob_loss:   2.55, total_loss:   5.20\n",
            " Total Loss  tf.Tensor(4.84717, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   71/125, lr:0.000001, giou_loss:   2.27, conf_loss:   0.14, prob_loss:   2.44, total_loss:   4.85\n",
            " Total Loss  tf.Tensor(3.8001044, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   72/125, lr:0.000001, giou_loss:   1.83, conf_loss:   0.06, prob_loss:   1.91, total_loss:   3.80\n",
            " Total Loss  tf.Tensor(4.4468746, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   73/125, lr:0.000001, giou_loss:   1.98, conf_loss:   0.14, prob_loss:   2.32, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(3.4200554, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   74/125, lr:0.000001, giou_loss:   1.51, conf_loss:   0.03, prob_loss:   1.87, total_loss:   3.42\n",
            " Total Loss  tf.Tensor(4.455034, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   75/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.03, prob_loss:   2.40, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(4.242403, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   76/125, lr:0.000001, giou_loss:   1.98, conf_loss:   0.04, prob_loss:   2.23, total_loss:   4.24\n",
            " Total Loss  tf.Tensor(4.348598, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   77/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.03, prob_loss:   2.38, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(4.850073, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   78/125, lr:0.000001, giou_loss:   2.37, conf_loss:   0.01, prob_loss:   2.47, total_loss:   4.85\n",
            " Total Loss  tf.Tensor(3.5337214, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   79/125, lr:0.000001, giou_loss:   1.48, conf_loss:   0.03, prob_loss:   2.03, total_loss:   3.53\n",
            " Total Loss  tf.Tensor(4.537496, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   80/125, lr:0.000001, giou_loss:   1.90, conf_loss:   0.27, prob_loss:   2.37, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(4.325853, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   81/125, lr:0.000001, giou_loss:   2.04, conf_loss:   0.03, prob_loss:   2.26, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(4.87237, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   82/125, lr:0.000001, giou_loss:   2.45, conf_loss:   0.03, prob_loss:   2.39, total_loss:   4.87\n",
            " Total Loss  tf.Tensor(4.640696, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   83/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.13, prob_loss:   2.30, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.640189, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   84/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.11, prob_loss:   2.46, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.942507, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   85/125, lr:0.000001, giou_loss:   2.04, conf_loss:   0.03, prob_loss:   2.87, total_loss:   4.94\n",
            " Total Loss  tf.Tensor(5.337533, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   86/125, lr:0.000001, giou_loss:   2.45, conf_loss:   0.30, prob_loss:   2.59, total_loss:   5.34\n",
            " Total Loss  tf.Tensor(4.438183, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   87/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.09, prob_loss:   2.38, total_loss:   4.44\n",
            " Total Loss  tf.Tensor(3.207924, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   88/125, lr:0.000001, giou_loss:   1.42, conf_loss:   0.01, prob_loss:   1.78, total_loss:   3.21\n",
            " Total Loss  tf.Tensor(4.5831537, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   89/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.02, prob_loss:   2.50, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.9620113, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   90/125, lr:0.000001, giou_loss:   2.29, conf_loss:   0.16, prob_loss:   2.51, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(4.019255, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   91/125, lr:0.000001, giou_loss:   1.78, conf_loss:   0.02, prob_loss:   2.21, total_loss:   4.02\n",
            " Total Loss  tf.Tensor(3.9463563, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   92/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.18, prob_loss:   1.96, total_loss:   3.95\n",
            " Total Loss  tf.Tensor(3.6641002, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   93/125, lr:0.000001, giou_loss:   1.47, conf_loss:   0.15, prob_loss:   2.04, total_loss:   3.66\n",
            " Total Loss  tf.Tensor(4.9883604, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   94/125, lr:0.000001, giou_loss:   2.42, conf_loss:   0.20, prob_loss:   2.38, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(3.702359, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   95/125, lr:0.000001, giou_loss:   1.55, conf_loss:   0.05, prob_loss:   2.10, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(2.9513803, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   96/125, lr:0.000001, giou_loss:   1.37, conf_loss:   0.01, prob_loss:   1.57, total_loss:   2.95\n",
            " Total Loss  tf.Tensor(5.973958, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   97/125, lr:0.000001, giou_loss:   2.87, conf_loss:   0.20, prob_loss:   2.91, total_loss:   5.97\n",
            " Total Loss  tf.Tensor(4.9018097, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   98/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.08, prob_loss:   2.50, total_loss:   4.90\n",
            " Total Loss  tf.Tensor(3.7926383, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:   99/125, lr:0.000001, giou_loss:   1.73, conf_loss:   0.17, prob_loss:   1.90, total_loss:   3.79\n",
            " Total Loss  tf.Tensor(4.524331, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  100/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.09, prob_loss:   2.22, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(4.17686, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  101/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.08, prob_loss:   2.17, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(4.362177, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  102/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.16, prob_loss:   2.20, total_loss:   4.36\n",
            " Total Loss  tf.Tensor(5.4239645, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  103/125, lr:0.000001, giou_loss:   2.42, conf_loss:   0.07, prob_loss:   2.93, total_loss:   5.42\n",
            " Total Loss  tf.Tensor(3.5612168, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  104/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.08, prob_loss:   1.82, total_loss:   3.56\n",
            " Total Loss  tf.Tensor(4.0722113, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  105/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.06, prob_loss:   2.19, total_loss:   4.07\n",
            " Total Loss  tf.Tensor(4.196768, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  106/125, lr:0.000001, giou_loss:   1.81, conf_loss:   0.02, prob_loss:   2.36, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(5.6559467, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  107/125, lr:0.000001, giou_loss:   2.61, conf_loss:   0.07, prob_loss:   2.98, total_loss:   5.66\n",
            " Total Loss  tf.Tensor(4.6734195, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  108/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.04, prob_loss:   2.53, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(3.2223158, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  109/125, lr:0.000001, giou_loss:   1.51, conf_loss:   0.07, prob_loss:   1.64, total_loss:   3.22\n",
            " Total Loss  tf.Tensor(4.6246996, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  110/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.03, prob_loss:   2.46, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(4.373231, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  111/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.18, prob_loss:   1.87, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(3.6434937, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  112/125, lr:0.000001, giou_loss:   1.81, conf_loss:   0.12, prob_loss:   1.72, total_loss:   3.64\n",
            " Total Loss  tf.Tensor(4.17266, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  113/125, lr:0.000001, giou_loss:   1.96, conf_loss:   0.21, prob_loss:   2.00, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.1655426, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  114/125, lr:0.000001, giou_loss:   1.86, conf_loss:   0.06, prob_loss:   2.25, total_loss:   4.17\n",
            " Total Loss  tf.Tensor(4.8586435, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  115/125, lr:0.000001, giou_loss:   2.31, conf_loss:   0.10, prob_loss:   2.45, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(4.9553523, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  116/125, lr:0.000001, giou_loss:   2.36, conf_loss:   0.11, prob_loss:   2.49, total_loss:   4.96\n",
            " Total Loss  tf.Tensor(4.588922, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  117/125, lr:0.000001, giou_loss:   2.01, conf_loss:   0.02, prob_loss:   2.56, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(5.432156, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  118/125, lr:0.000001, giou_loss:   2.60, conf_loss:   0.08, prob_loss:   2.75, total_loss:   5.43\n",
            " Total Loss  tf.Tensor(4.0288677, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  119/125, lr:0.000001, giou_loss:   1.91, conf_loss:   0.04, prob_loss:   2.08, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(6.916751, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  120/125, lr:0.000001, giou_loss:   2.81, conf_loss:   1.48, prob_loss:   2.62, total_loss:   6.92\n",
            " Total Loss  tf.Tensor(4.4593678, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  121/125, lr:0.000001, giou_loss:   2.08, conf_loss:   0.12, prob_loss:   2.26, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(3.0674465, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  122/125, lr:0.000001, giou_loss:   1.35, conf_loss:   0.04, prob_loss:   1.67, total_loss:   3.07\n",
            " Total Loss  tf.Tensor(4.3069334, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  123/125, lr:0.000001, giou_loss:   1.92, conf_loss:   0.04, prob_loss:   2.34, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.4128866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:  124/125, lr:0.000001, giou_loss:   2.34, conf_loss:   0.09, prob_loss:   1.98, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(4.8830986, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    0/125, lr:0.000001, giou_loss:   2.38, conf_loss:   0.02, prob_loss:   2.48, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(3.7953424, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:97 step:    1/125, lr:0.000001, giou_loss:   1.53, conf_loss:   0.39, prob_loss:   1.88, total_loss:   3.80\n",
            "\n",
            "\n",
            "giou_val_loss:   2.20, conf_val_loss:   0.62, prob_val_loss:   5.29, total_val_loss:   8.12\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.4484057, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    2/125, lr:0.000001, giou_loss:   1.96, conf_loss:   0.03, prob_loss:   2.46, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(5.1261125, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    3/125, lr:0.000001, giou_loss:   2.46, conf_loss:   0.11, prob_loss:   2.55, total_loss:   5.13\n",
            " Total Loss  tf.Tensor(4.6338797, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    4/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.23, prob_loss:   2.34, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(3.2487242, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    5/125, lr:0.000001, giou_loss:   1.56, conf_loss:   0.02, prob_loss:   1.67, total_loss:   3.25\n",
            " Total Loss  tf.Tensor(4.516061, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    6/125, lr:0.000001, giou_loss:   2.18, conf_loss:   0.03, prob_loss:   2.31, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(4.968174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    7/125, lr:0.000001, giou_loss:   2.35, conf_loss:   0.05, prob_loss:   2.57, total_loss:   4.97\n",
            " Total Loss  tf.Tensor(5.150161, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    8/125, lr:0.000001, giou_loss:   2.41, conf_loss:   0.32, prob_loss:   2.42, total_loss:   5.15\n",
            " Total Loss  tf.Tensor(3.6553812, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    9/125, lr:0.000001, giou_loss:   1.59, conf_loss:   0.07, prob_loss:   1.99, total_loss:   3.66\n",
            " Total Loss  tf.Tensor(4.1286325, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   10/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.05, prob_loss:   1.95, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(5.6115794, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   11/125, lr:0.000001, giou_loss:   2.61, conf_loss:   0.25, prob_loss:   2.75, total_loss:   5.61\n",
            " Total Loss  tf.Tensor(4.776382, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   12/125, lr:0.000001, giou_loss:   2.23, conf_loss:   0.07, prob_loss:   2.48, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(4.9089813, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   13/125, lr:0.000001, giou_loss:   2.15, conf_loss:   0.03, prob_loss:   2.72, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(3.762708, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   14/125, lr:0.000001, giou_loss:   1.70, conf_loss:   0.06, prob_loss:   2.00, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(5.4880333, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   15/125, lr:0.000001, giou_loss:   2.43, conf_loss:   0.14, prob_loss:   2.92, total_loss:   5.49\n",
            " Total Loss  tf.Tensor(4.8759985, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   16/125, lr:0.000001, giou_loss:   2.29, conf_loss:   0.21, prob_loss:   2.38, total_loss:   4.88\n",
            " Total Loss  tf.Tensor(4.5585375, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   17/125, lr:0.000001, giou_loss:   1.91, conf_loss:   0.05, prob_loss:   2.60, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.523069, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   18/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.10, prob_loss:   2.49, total_loss:   4.52\n",
            " Total Loss  tf.Tensor(4.6321297, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   19/125, lr:0.000001, giou_loss:   2.16, conf_loss:   0.02, prob_loss:   2.45, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(4.299095, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   20/125, lr:0.000001, giou_loss:   1.95, conf_loss:   0.07, prob_loss:   2.28, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(4.766453, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   21/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.09, prob_loss:   2.44, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(3.8542929, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   22/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.01, prob_loss:   1.86, total_loss:   3.85\n",
            " Total Loss  tf.Tensor(4.862444, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   23/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.12, prob_loss:   2.50, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(3.7383504, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   24/125, lr:0.000001, giou_loss:   1.88, conf_loss:   0.05, prob_loss:   1.81, total_loss:   3.74\n",
            " Total Loss  tf.Tensor(4.8298354, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   25/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.03, prob_loss:   2.59, total_loss:   4.83\n",
            " Total Loss  tf.Tensor(3.925201, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   26/125, lr:0.000001, giou_loss:   1.73, conf_loss:   0.04, prob_loss:   2.16, total_loss:   3.93\n",
            " Total Loss  tf.Tensor(4.576027, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   27/125, lr:0.000001, giou_loss:   2.09, conf_loss:   0.02, prob_loss:   2.47, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(4.591055, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   28/125, lr:0.000001, giou_loss:   2.31, conf_loss:   0.04, prob_loss:   2.25, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(5.9417043, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   29/125, lr:0.000001, giou_loss:   2.91, conf_loss:   0.08, prob_loss:   2.96, total_loss:   5.94\n",
            " Total Loss  tf.Tensor(3.7279248, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   30/125, lr:0.000001, giou_loss:   1.61, conf_loss:   0.10, prob_loss:   2.01, total_loss:   3.73\n",
            " Total Loss  tf.Tensor(3.9370346, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   31/125, lr:0.000001, giou_loss:   1.73, conf_loss:   0.03, prob_loss:   2.18, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(3.4673655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   32/125, lr:0.000001, giou_loss:   1.64, conf_loss:   0.01, prob_loss:   1.82, total_loss:   3.47\n",
            " Total Loss  tf.Tensor(5.1303043, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   33/125, lr:0.000001, giou_loss:   2.38, conf_loss:   0.13, prob_loss:   2.61, total_loss:   5.13\n",
            " Total Loss  tf.Tensor(4.646549, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   34/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.24, prob_loss:   2.27, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(4.8013716, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   35/125, lr:0.000001, giou_loss:   2.16, conf_loss:   0.09, prob_loss:   2.55, total_loss:   4.80\n",
            " Total Loss  tf.Tensor(5.871536, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   36/125, lr:0.000001, giou_loss:   2.80, conf_loss:   0.14, prob_loss:   2.93, total_loss:   5.87\n",
            " Total Loss  tf.Tensor(4.2536354, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   37/125, lr:0.000001, giou_loss:   1.90, conf_loss:   0.12, prob_loss:   2.23, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(5.018014, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   38/125, lr:0.000001, giou_loss:   2.22, conf_loss:   0.17, prob_loss:   2.63, total_loss:   5.02\n",
            " Total Loss  tf.Tensor(4.6319656, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   39/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.18, prob_loss:   2.33, total_loss:   4.63\n",
            " Total Loss  tf.Tensor(2.97665, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   40/125, lr:0.000001, giou_loss:   1.44, conf_loss:   0.04, prob_loss:   1.50, total_loss:   2.98\n",
            " Total Loss  tf.Tensor(4.4032836, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   41/125, lr:0.000001, giou_loss:   2.01, conf_loss:   0.16, prob_loss:   2.23, total_loss:   4.40\n",
            " Total Loss  tf.Tensor(3.6953187, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   42/125, lr:0.000001, giou_loss:   1.67, conf_loss:   0.02, prob_loss:   2.01, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(4.6217756, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   43/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.07, prob_loss:   2.48, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(4.1446524, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   44/125, lr:0.000001, giou_loss:   1.85, conf_loss:   0.04, prob_loss:   2.26, total_loss:   4.14\n",
            " Total Loss  tf.Tensor(4.3740664, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   45/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.06, prob_loss:   2.19, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(4.294346, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   46/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.04, prob_loss:   2.18, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(3.909271, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   47/125, lr:0.000001, giou_loss:   1.78, conf_loss:   0.04, prob_loss:   2.09, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(5.884849, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   48/125, lr:0.000001, giou_loss:   2.80, conf_loss:   0.11, prob_loss:   2.97, total_loss:   5.88\n",
            " Total Loss  tf.Tensor(4.1553473, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   49/125, lr:0.000001, giou_loss:   2.20, conf_loss:   0.03, prob_loss:   1.93, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(5.2122526, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   50/125, lr:0.000001, giou_loss:   2.40, conf_loss:   0.08, prob_loss:   2.72, total_loss:   5.21\n",
            " Total Loss  tf.Tensor(4.513978, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   51/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.23, prob_loss:   2.26, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.64666, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   52/125, lr:0.000001, giou_loss:   2.26, conf_loss:   0.02, prob_loss:   2.36, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(3.8895764, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   53/125, lr:0.000001, giou_loss:   1.77, conf_loss:   0.08, prob_loss:   2.05, total_loss:   3.89\n",
            " Total Loss  tf.Tensor(4.3140535, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   54/125, lr:0.000001, giou_loss:   1.86, conf_loss:   0.19, prob_loss:   2.26, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(4.987031, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   55/125, lr:0.000001, giou_loss:   2.30, conf_loss:   0.07, prob_loss:   2.62, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(4.6176224, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   56/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.11, prob_loss:   2.46, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(5.026468, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   57/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.01, prob_loss:   2.77, total_loss:   5.03\n",
            " Total Loss  tf.Tensor(5.29667, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   58/125, lr:0.000001, giou_loss:   2.38, conf_loss:   0.20, prob_loss:   2.72, total_loss:   5.30\n",
            " Total Loss  tf.Tensor(4.7067375, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   59/125, lr:0.000001, giou_loss:   2.20, conf_loss:   0.06, prob_loss:   2.45, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(5.0212326, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   60/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.28, prob_loss:   2.61, total_loss:   5.02\n",
            " Total Loss  tf.Tensor(5.782648, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   61/125, lr:0.000001, giou_loss:   2.74, conf_loss:   0.18, prob_loss:   2.87, total_loss:   5.78\n",
            " Total Loss  tf.Tensor(3.0217962, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   62/125, lr:0.000001, giou_loss:   1.32, conf_loss:   0.08, prob_loss:   1.63, total_loss:   3.02\n",
            " Total Loss  tf.Tensor(4.7009077, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   63/125, lr:0.000001, giou_loss:   2.26, conf_loss:   0.07, prob_loss:   2.37, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(3.71416, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   64/125, lr:0.000001, giou_loss:   1.61, conf_loss:   0.07, prob_loss:   2.03, total_loss:   3.71\n",
            " Total Loss  tf.Tensor(5.746215, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   65/125, lr:0.000001, giou_loss:   2.56, conf_loss:   0.07, prob_loss:   3.12, total_loss:   5.75\n",
            " Total Loss  tf.Tensor(4.203866, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   66/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.18, prob_loss:   1.97, total_loss:   4.20\n",
            " Total Loss  tf.Tensor(5.3689995, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   67/125, lr:0.000001, giou_loss:   2.39, conf_loss:   0.13, prob_loss:   2.84, total_loss:   5.37\n",
            " Total Loss  tf.Tensor(3.9243448, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   68/125, lr:0.000001, giou_loss:   1.98, conf_loss:   0.10, prob_loss:   1.85, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(4.9156485, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   69/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.02, prob_loss:   2.61, total_loss:   4.92\n",
            " Total Loss  tf.Tensor(4.5605345, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   70/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.15, prob_loss:   2.28, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.4591894, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   71/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.08, prob_loss:   2.35, total_loss:   4.46\n",
            " Total Loss  tf.Tensor(4.5327244, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   72/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.07, prob_loss:   2.39, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(4.845616, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   73/125, lr:0.000001, giou_loss:   2.19, conf_loss:   0.39, prob_loss:   2.27, total_loss:   4.85\n",
            " Total Loss  tf.Tensor(3.784442, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   74/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.06, prob_loss:   1.92, total_loss:   3.78\n",
            " Total Loss  tf.Tensor(4.773945, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   75/125, lr:0.000001, giou_loss:   2.34, conf_loss:   0.04, prob_loss:   2.40, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(4.0049953, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   76/125, lr:0.000001, giou_loss:   1.79, conf_loss:   0.03, prob_loss:   2.19, total_loss:   4.00\n",
            " Total Loss  tf.Tensor(3.4698434, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   77/125, lr:0.000001, giou_loss:   1.51, conf_loss:   0.01, prob_loss:   1.95, total_loss:   3.47\n",
            " Total Loss  tf.Tensor(3.4671617, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   78/125, lr:0.000001, giou_loss:   1.59, conf_loss:   0.01, prob_loss:   1.86, total_loss:   3.47\n",
            " Total Loss  tf.Tensor(3.2295632, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   79/125, lr:0.000001, giou_loss:   1.47, conf_loss:   0.16, prob_loss:   1.59, total_loss:   3.23\n",
            " Total Loss  tf.Tensor(4.743062, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   80/125, lr:0.000001, giou_loss:   2.44, conf_loss:   0.39, prob_loss:   1.91, total_loss:   4.74\n",
            " Total Loss  tf.Tensor(5.6327424, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   81/125, lr:0.000001, giou_loss:   2.68, conf_loss:   0.20, prob_loss:   2.75, total_loss:   5.63\n",
            " Total Loss  tf.Tensor(4.782256, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   82/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.02, prob_loss:   2.48, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(3.709134, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   83/125, lr:0.000001, giou_loss:   1.59, conf_loss:   0.02, prob_loss:   2.10, total_loss:   3.71\n",
            " Total Loss  tf.Tensor(4.1303043, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   84/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.08, prob_loss:   2.23, total_loss:   4.13\n",
            " Total Loss  tf.Tensor(4.588191, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   85/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.05, prob_loss:   2.33, total_loss:   4.59\n",
            " Total Loss  tf.Tensor(3.8657157, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   86/125, lr:0.000001, giou_loss:   1.65, conf_loss:   0.01, prob_loss:   2.20, total_loss:   3.87\n",
            " Total Loss  tf.Tensor(5.1226244, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   87/125, lr:0.000001, giou_loss:   2.39, conf_loss:   0.04, prob_loss:   2.70, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(3.4555054, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   88/125, lr:0.000001, giou_loss:   1.43, conf_loss:   0.01, prob_loss:   2.02, total_loss:   3.46\n",
            " Total Loss  tf.Tensor(4.6969967, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   89/125, lr:0.000001, giou_loss:   2.18, conf_loss:   0.20, prob_loss:   2.32, total_loss:   4.70\n",
            " Total Loss  tf.Tensor(5.4008904, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   90/125, lr:0.000001, giou_loss:   2.78, conf_loss:   0.18, prob_loss:   2.44, total_loss:   5.40\n",
            " Total Loss  tf.Tensor(3.8191462, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   91/125, lr:0.000001, giou_loss:   1.72, conf_loss:   0.18, prob_loss:   1.92, total_loss:   3.82\n",
            " Total Loss  tf.Tensor(3.9612832, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   92/125, lr:0.000001, giou_loss:   1.95, conf_loss:   0.04, prob_loss:   1.97, total_loss:   3.96\n",
            " Total Loss  tf.Tensor(4.7354383, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   93/125, lr:0.000001, giou_loss:   1.91, conf_loss:   0.43, prob_loss:   2.39, total_loss:   4.74\n",
            " Total Loss  tf.Tensor(3.9713242, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   94/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.05, prob_loss:   1.94, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(3.741221, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   95/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.02, prob_loss:   2.06, total_loss:   3.74\n",
            " Total Loss  tf.Tensor(4.209401, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   96/125, lr:0.000001, giou_loss:   1.83, conf_loss:   0.04, prob_loss:   2.34, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.286794, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   97/125, lr:0.000001, giou_loss:   1.77, conf_loss:   0.08, prob_loss:   2.43, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(5.109186, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   98/125, lr:0.000001, giou_loss:   2.33, conf_loss:   0.06, prob_loss:   2.72, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(4.503193, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:   99/125, lr:0.000001, giou_loss:   2.04, conf_loss:   0.18, prob_loss:   2.28, total_loss:   4.50\n",
            " Total Loss  tf.Tensor(4.9118814, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  100/125, lr:0.000001, giou_loss:   2.27, conf_loss:   0.07, prob_loss:   2.58, total_loss:   4.91\n",
            " Total Loss  tf.Tensor(3.3571172, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  101/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.06, prob_loss:   1.64, total_loss:   3.36\n",
            " Total Loss  tf.Tensor(4.7469406, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  102/125, lr:0.000001, giou_loss:   2.35, conf_loss:   0.16, prob_loss:   2.24, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(6.012918, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  103/125, lr:0.000001, giou_loss:   2.72, conf_loss:   0.23, prob_loss:   3.07, total_loss:   6.01\n",
            " Total Loss  tf.Tensor(5.2867584, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  104/125, lr:0.000001, giou_loss:   2.63, conf_loss:   0.02, prob_loss:   2.63, total_loss:   5.29\n",
            " Total Loss  tf.Tensor(4.2071, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  105/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.06, prob_loss:   2.16, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(5.170057, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  106/125, lr:0.000001, giou_loss:   2.33, conf_loss:   0.09, prob_loss:   2.75, total_loss:   5.17\n",
            " Total Loss  tf.Tensor(3.2587547, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  107/125, lr:0.000001, giou_loss:   1.34, conf_loss:   0.01, prob_loss:   1.90, total_loss:   3.26\n",
            " Total Loss  tf.Tensor(4.1079884, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  108/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.18, prob_loss:   1.93, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(5.2299566, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  109/125, lr:0.000001, giou_loss:   2.39, conf_loss:   0.07, prob_loss:   2.77, total_loss:   5.23\n",
            " Total Loss  tf.Tensor(3.3966599, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  110/125, lr:0.000001, giou_loss:   1.59, conf_loss:   0.02, prob_loss:   1.79, total_loss:   3.40\n",
            " Total Loss  tf.Tensor(4.3305254, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  111/125, lr:0.000001, giou_loss:   1.89, conf_loss:   0.05, prob_loss:   2.39, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(4.026738, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  112/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.06, prob_loss:   2.23, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(4.886861, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  113/125, lr:0.000001, giou_loss:   2.27, conf_loss:   0.04, prob_loss:   2.58, total_loss:   4.89\n",
            " Total Loss  tf.Tensor(4.747055, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  114/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.02, prob_loss:   2.60, total_loss:   4.75\n",
            " Total Loss  tf.Tensor(4.162115, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  115/125, lr:0.000001, giou_loss:   1.92, conf_loss:   0.06, prob_loss:   2.19, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(4.707261, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  116/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.11, prob_loss:   2.48, total_loss:   4.71\n",
            " Total Loss  tf.Tensor(5.0650644, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  117/125, lr:0.000001, giou_loss:   2.44, conf_loss:   0.11, prob_loss:   2.51, total_loss:   5.07\n",
            " Total Loss  tf.Tensor(4.1146317, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  118/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.02, prob_loss:   2.12, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(5.6783657, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  119/125, lr:0.000001, giou_loss:   2.54, conf_loss:   0.22, prob_loss:   2.91, total_loss:   5.68\n",
            " Total Loss  tf.Tensor(4.5483174, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  120/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.04, prob_loss:   2.51, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(5.619643, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  121/125, lr:0.000001, giou_loss:   2.58, conf_loss:   0.25, prob_loss:   2.79, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(5.5376663, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  122/125, lr:0.000001, giou_loss:   2.58, conf_loss:   0.25, prob_loss:   2.71, total_loss:   5.54\n",
            " Total Loss  tf.Tensor(5.1438384, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  123/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.03, prob_loss:   2.83, total_loss:   5.14\n",
            " Total Loss  tf.Tensor(4.2425036, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:  124/125, lr:0.000001, giou_loss:   1.77, conf_loss:   0.41, prob_loss:   2.06, total_loss:   4.24\n",
            " Total Loss  tf.Tensor(4.247305, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    0/125, lr:0.000001, giou_loss:   2.11, conf_loss:   0.11, prob_loss:   2.03, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(6.6174183, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:98 step:    1/125, lr:0.000001, giou_loss:   2.50, conf_loss:   1.41, prob_loss:   2.71, total_loss:   6.62\n",
            "\n",
            "\n",
            "giou_val_loss:   2.21, conf_val_loss:   0.65, prob_val_loss:   5.30, total_val_loss:   8.17\n",
            "\n",
            "\n",
            " Total Loss  tf.Tensor(4.7907867, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    2/125, lr:0.000001, giou_loss:   2.30, conf_loss:   0.09, prob_loss:   2.40, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(4.6432533, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    3/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.04, prob_loss:   2.32, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.424292, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    4/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.10, prob_loss:   2.30, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(4.354289, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    5/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.01, prob_loss:   2.29, total_loss:   4.35\n",
            " Total Loss  tf.Tensor(4.416111, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    6/125, lr:0.000001, giou_loss:   1.92, conf_loss:   0.19, prob_loss:   2.31, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(4.1500845, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    7/125, lr:0.000001, giou_loss:   1.96, conf_loss:   0.10, prob_loss:   2.09, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(4.1540565, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    8/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.01, prob_loss:   2.34, total_loss:   4.15\n",
            " Total Loss  tf.Tensor(3.6960783, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    9/125, lr:0.000001, giou_loss:   1.74, conf_loss:   0.14, prob_loss:   1.81, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(4.2092857, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   10/125, lr:0.000001, giou_loss:   2.12, conf_loss:   0.05, prob_loss:   2.04, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(3.8494825, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   11/125, lr:0.000001, giou_loss:   1.70, conf_loss:   0.03, prob_loss:   2.11, total_loss:   3.85\n",
            " Total Loss  tf.Tensor(3.4845808, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   12/125, lr:0.000001, giou_loss:   1.56, conf_loss:   0.03, prob_loss:   1.89, total_loss:   3.48\n",
            " Total Loss  tf.Tensor(3.7559876, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   13/125, lr:0.000001, giou_loss:   1.56, conf_loss:   0.05, prob_loss:   2.15, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(4.4474697, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   14/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.09, prob_loss:   2.30, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.50803, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   15/125, lr:0.000001, giou_loss:   2.22, conf_loss:   0.02, prob_loss:   2.26, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(3.1356995, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   16/125, lr:0.000001, giou_loss:   1.41, conf_loss:   0.04, prob_loss:   1.69, total_loss:   3.14\n",
            " Total Loss  tf.Tensor(5.2381577, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   17/125, lr:0.000001, giou_loss:   2.32, conf_loss:   0.07, prob_loss:   2.85, total_loss:   5.24\n",
            " Total Loss  tf.Tensor(4.505747, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   18/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.10, prob_loss:   2.19, total_loss:   4.51\n",
            " Total Loss  tf.Tensor(4.9733777, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   19/125, lr:0.000001, giou_loss:   2.21, conf_loss:   0.13, prob_loss:   2.63, total_loss:   4.97\n",
            " Total Loss  tf.Tensor(3.4231932, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   20/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.06, prob_loss:   1.61, total_loss:   3.42\n",
            " Total Loss  tf.Tensor(3.9737515, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   21/125, lr:0.000001, giou_loss:   1.77, conf_loss:   0.03, prob_loss:   2.18, total_loss:   3.97\n",
            " Total Loss  tf.Tensor(4.1847944, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   22/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.06, prob_loss:   2.13, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(4.3108916, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   23/125, lr:0.000001, giou_loss:   1.78, conf_loss:   0.19, prob_loss:   2.34, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(3.8432665, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   24/125, lr:0.000001, giou_loss:   1.66, conf_loss:   0.10, prob_loss:   2.08, total_loss:   3.84\n",
            " Total Loss  tf.Tensor(5.4593015, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   25/125, lr:0.000001, giou_loss:   2.50, conf_loss:   0.08, prob_loss:   2.88, total_loss:   5.46\n",
            " Total Loss  tf.Tensor(4.1828256, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   26/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.02, prob_loss:   2.35, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(5.6159925, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   27/125, lr:0.000001, giou_loss:   2.64, conf_loss:   0.05, prob_loss:   2.93, total_loss:   5.62\n",
            " Total Loss  tf.Tensor(3.2204056, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   28/125, lr:0.000001, giou_loss:   1.59, conf_loss:   0.03, prob_loss:   1.61, total_loss:   3.22\n",
            " Total Loss  tf.Tensor(4.5362663, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   29/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.06, prob_loss:   2.38, total_loss:   4.54\n",
            " Total Loss  tf.Tensor(4.7802553, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   30/125, lr:0.000001, giou_loss:   2.41, conf_loss:   0.04, prob_loss:   2.33, total_loss:   4.78\n",
            " Total Loss  tf.Tensor(3.9938147, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   31/125, lr:0.000001, giou_loss:   1.86, conf_loss:   0.12, prob_loss:   2.02, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(4.4082117, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   32/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.04, prob_loss:   2.34, total_loss:   4.41\n",
            " Total Loss  tf.Tensor(3.9202907, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   33/125, lr:0.000001, giou_loss:   1.75, conf_loss:   0.13, prob_loss:   2.04, total_loss:   3.92\n",
            " Total Loss  tf.Tensor(3.667027, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   34/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.01, prob_loss:   1.85, total_loss:   3.67\n",
            " Total Loss  tf.Tensor(6.406069, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   35/125, lr:0.000001, giou_loss:   2.77, conf_loss:   0.35, prob_loss:   3.28, total_loss:   6.41\n",
            " Total Loss  tf.Tensor(4.4680367, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   36/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.08, prob_loss:   2.24, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(3.2290485, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   37/125, lr:0.000001, giou_loss:   1.54, conf_loss:   0.02, prob_loss:   1.67, total_loss:   3.23\n",
            " Total Loss  tf.Tensor(4.863599, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   38/125, lr:0.000001, giou_loss:   2.22, conf_loss:   0.13, prob_loss:   2.51, total_loss:   4.86\n",
            " Total Loss  tf.Tensor(4.549959, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   39/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.22, prob_loss:   2.28, total_loss:   4.55\n",
            " Total Loss  tf.Tensor(3.7646031, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   40/125, lr:0.000001, giou_loss:   1.71, conf_loss:   0.04, prob_loss:   2.01, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(4.179395, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   41/125, lr:0.000001, giou_loss:   1.77, conf_loss:   0.04, prob_loss:   2.37, total_loss:   4.18\n",
            " Total Loss  tf.Tensor(4.262706, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   42/125, lr:0.000001, giou_loss:   1.95, conf_loss:   0.07, prob_loss:   2.25, total_loss:   4.26\n",
            " Total Loss  tf.Tensor(4.2555714, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   43/125, lr:0.000001, giou_loss:   2.01, conf_loss:   0.18, prob_loss:   2.07, total_loss:   4.26\n",
            " Total Loss  tf.Tensor(5.7583838, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   44/125, lr:0.000001, giou_loss:   2.62, conf_loss:   0.25, prob_loss:   2.89, total_loss:   5.76\n",
            " Total Loss  tf.Tensor(5.5619144, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   45/125, lr:0.000001, giou_loss:   2.63, conf_loss:   0.17, prob_loss:   2.76, total_loss:   5.56\n",
            " Total Loss  tf.Tensor(4.669691, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   46/125, lr:0.000001, giou_loss:   2.09, conf_loss:   0.07, prob_loss:   2.51, total_loss:   4.67\n",
            " Total Loss  tf.Tensor(5.2922506, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   47/125, lr:0.000001, giou_loss:   2.64, conf_loss:   0.13, prob_loss:   2.51, total_loss:   5.29\n",
            " Total Loss  tf.Tensor(3.3092241, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   48/125, lr:0.000001, giou_loss:   1.67, conf_loss:   0.01, prob_loss:   1.63, total_loss:   3.31\n",
            " Total Loss  tf.Tensor(3.990954, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   49/125, lr:0.000001, giou_loss:   1.89, conf_loss:   0.05, prob_loss:   2.05, total_loss:   3.99\n",
            " Total Loss  tf.Tensor(3.4673667, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   50/125, lr:0.000001, giou_loss:   1.51, conf_loss:   0.07, prob_loss:   1.88, total_loss:   3.47\n",
            " Total Loss  tf.Tensor(3.7634563, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   51/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.01, prob_loss:   1.95, total_loss:   3.76\n",
            " Total Loss  tf.Tensor(5.8905473, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   52/125, lr:0.000001, giou_loss:   2.86, conf_loss:   0.25, prob_loss:   2.78, total_loss:   5.89\n",
            " Total Loss  tf.Tensor(3.9144478, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   53/125, lr:0.000001, giou_loss:   1.91, conf_loss:   0.13, prob_loss:   1.87, total_loss:   3.91\n",
            " Total Loss  tf.Tensor(4.334805, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   54/125, lr:0.000001, giou_loss:   2.09, conf_loss:   0.10, prob_loss:   2.15, total_loss:   4.33\n",
            " Total Loss  tf.Tensor(4.1945744, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   55/125, lr:0.000001, giou_loss:   1.86, conf_loss:   0.10, prob_loss:   2.24, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(4.6453023, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   56/125, lr:0.000001, giou_loss:   1.97, conf_loss:   0.11, prob_loss:   2.57, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(4.7908955, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   57/125, lr:0.000001, giou_loss:   2.24, conf_loss:   0.09, prob_loss:   2.46, total_loss:   4.79\n",
            " Total Loss  tf.Tensor(3.6164937, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   58/125, lr:0.000001, giou_loss:   1.53, conf_loss:   0.23, prob_loss:   1.86, total_loss:   3.62\n",
            " Total Loss  tf.Tensor(5.112075, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   59/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.38, prob_loss:   2.46, total_loss:   5.11\n",
            " Total Loss  tf.Tensor(3.4043655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   60/125, lr:0.000001, giou_loss:   1.38, conf_loss:   0.01, prob_loss:   2.02, total_loss:   3.40\n",
            " Total Loss  tf.Tensor(3.7090945, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   61/125, lr:0.000001, giou_loss:   1.60, conf_loss:   0.01, prob_loss:   2.10, total_loss:   3.71\n",
            " Total Loss  tf.Tensor(5.690751, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   62/125, lr:0.000001, giou_loss:   2.57, conf_loss:   0.09, prob_loss:   3.02, total_loss:   5.69\n",
            " Total Loss  tf.Tensor(4.3074017, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   63/125, lr:0.000001, giou_loss:   1.92, conf_loss:   0.07, prob_loss:   2.32, total_loss:   4.31\n",
            " Total Loss  tf.Tensor(3.9433608, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   64/125, lr:0.000001, giou_loss:   1.68, conf_loss:   0.03, prob_loss:   2.24, total_loss:   3.94\n",
            " Total Loss  tf.Tensor(4.6618185, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   65/125, lr:0.000001, giou_loss:   2.19, conf_loss:   0.05, prob_loss:   2.42, total_loss:   4.66\n",
            " Total Loss  tf.Tensor(4.0496297, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   66/125, lr:0.000001, giou_loss:   1.83, conf_loss:   0.02, prob_loss:   2.20, total_loss:   4.05\n",
            " Total Loss  tf.Tensor(4.465662, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   67/125, lr:0.000001, giou_loss:   2.19, conf_loss:   0.08, prob_loss:   2.20, total_loss:   4.47\n",
            " Total Loss  tf.Tensor(4.640028, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   68/125, lr:0.000001, giou_loss:   2.09, conf_loss:   0.11, prob_loss:   2.45, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.4345636, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   69/125, lr:0.000001, giou_loss:   1.96, conf_loss:   0.17, prob_loss:   2.31, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(3.4295142, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   70/125, lr:0.000001, giou_loss:   1.63, conf_loss:   0.02, prob_loss:   1.79, total_loss:   3.43\n",
            " Total Loss  tf.Tensor(4.426409, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   71/125, lr:0.000001, giou_loss:   2.06, conf_loss:   0.09, prob_loss:   2.28, total_loss:   4.43\n",
            " Total Loss  tf.Tensor(5.0333605, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   72/125, lr:0.000001, giou_loss:   2.23, conf_loss:   0.10, prob_loss:   2.71, total_loss:   5.03\n",
            " Total Loss  tf.Tensor(4.034815, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   73/125, lr:0.000001, giou_loss:   1.76, conf_loss:   0.01, prob_loss:   2.26, total_loss:   4.03\n",
            " Total Loss  tf.Tensor(4.562359, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   74/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.04, prob_loss:   2.43, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.5265036, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   75/125, lr:0.000001, giou_loss:   2.01, conf_loss:   0.15, prob_loss:   2.36, total_loss:   4.53\n",
            " Total Loss  tf.Tensor(4.4198933, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   76/125, lr:0.000001, giou_loss:   1.82, conf_loss:   0.05, prob_loss:   2.55, total_loss:   4.42\n",
            " Total Loss  tf.Tensor(4.98579, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   77/125, lr:0.000001, giou_loss:   2.28, conf_loss:   0.05, prob_loss:   2.66, total_loss:   4.99\n",
            " Total Loss  tf.Tensor(5.1226215, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   78/125, lr:0.000001, giou_loss:   2.33, conf_loss:   0.29, prob_loss:   2.50, total_loss:   5.12\n",
            " Total Loss  tf.Tensor(5.5620766, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   79/125, lr:0.000001, giou_loss:   2.56, conf_loss:   0.11, prob_loss:   2.89, total_loss:   5.56\n",
            " Total Loss  tf.Tensor(4.564167, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   80/125, lr:0.000001, giou_loss:   2.07, conf_loss:   0.03, prob_loss:   2.46, total_loss:   4.56\n",
            " Total Loss  tf.Tensor(4.723732, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   81/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.17, prob_loss:   2.50, total_loss:   4.72\n",
            " Total Loss  tf.Tensor(5.0481305, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   82/125, lr:0.000001, giou_loss:   2.16, conf_loss:   0.03, prob_loss:   2.86, total_loss:   5.05\n",
            " Total Loss  tf.Tensor(3.6748106, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   83/125, lr:0.000001, giou_loss:   1.68, conf_loss:   0.03, prob_loss:   1.97, total_loss:   3.67\n",
            " Total Loss  tf.Tensor(4.606478, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   84/125, lr:0.000001, giou_loss:   2.43, conf_loss:   0.06, prob_loss:   2.12, total_loss:   4.61\n",
            " Total Loss  tf.Tensor(4.3047795, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   85/125, lr:0.000001, giou_loss:   1.93, conf_loss:   0.17, prob_loss:   2.20, total_loss:   4.30\n",
            " Total Loss  tf.Tensor(5.3640423, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   86/125, lr:0.000001, giou_loss:   2.49, conf_loss:   0.05, prob_loss:   2.82, total_loss:   5.36\n",
            " Total Loss  tf.Tensor(5.318184, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   87/125, lr:0.000001, giou_loss:   2.51, conf_loss:   0.02, prob_loss:   2.79, total_loss:   5.32\n",
            " Total Loss  tf.Tensor(4.44948, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   88/125, lr:0.000001, giou_loss:   2.02, conf_loss:   0.03, prob_loss:   2.40, total_loss:   4.45\n",
            " Total Loss  tf.Tensor(4.1059585, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   89/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.02, prob_loss:   2.10, total_loss:   4.11\n",
            " Total Loss  tf.Tensor(3.0618386, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   90/125, lr:0.000001, giou_loss:   1.41, conf_loss:   0.02, prob_loss:   1.63, total_loss:   3.06\n",
            " Total Loss  tf.Tensor(4.2367973, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   91/125, lr:0.000001, giou_loss:   2.03, conf_loss:   0.07, prob_loss:   2.14, total_loss:   4.24\n",
            " Total Loss  tf.Tensor(4.2086844, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   92/125, lr:0.000001, giou_loss:   1.84, conf_loss:   0.11, prob_loss:   2.26, total_loss:   4.21\n",
            " Total Loss  tf.Tensor(4.1929655, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   93/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.10, prob_loss:   2.15, total_loss:   4.19\n",
            " Total Loss  tf.Tensor(5.6691217, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   94/125, lr:0.000001, giou_loss:   2.53, conf_loss:   0.28, prob_loss:   2.86, total_loss:   5.67\n",
            " Total Loss  tf.Tensor(4.485438, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   95/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.07, prob_loss:   2.48, total_loss:   4.49\n",
            " Total Loss  tf.Tensor(4.759338, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   96/125, lr:0.000001, giou_loss:   2.17, conf_loss:   0.05, prob_loss:   2.53, total_loss:   4.76\n",
            " Total Loss  tf.Tensor(5.275419, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   97/125, lr:0.000001, giou_loss:   2.53, conf_loss:   0.09, prob_loss:   2.66, total_loss:   5.28\n",
            " Total Loss  tf.Tensor(5.0959606, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   98/125, lr:0.000001, giou_loss:   2.26, conf_loss:   0.15, prob_loss:   2.69, total_loss:   5.10\n",
            " Total Loss  tf.Tensor(3.4586847, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:   99/125, lr:0.000001, giou_loss:   1.78, conf_loss:   0.07, prob_loss:   1.61, total_loss:   3.46\n",
            " Total Loss  tf.Tensor(4.64509, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  100/125, lr:0.000001, giou_loss:   2.27, conf_loss:   0.15, prob_loss:   2.23, total_loss:   4.65\n",
            " Total Loss  tf.Tensor(3.2767687, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  101/125, lr:0.000001, giou_loss:   1.52, conf_loss:   0.04, prob_loss:   1.72, total_loss:   3.28\n",
            " Total Loss  tf.Tensor(5.0809984, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  102/125, lr:0.000001, giou_loss:   2.43, conf_loss:   0.06, prob_loss:   2.59, total_loss:   5.08\n",
            " Total Loss  tf.Tensor(5.0308514, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  103/125, lr:0.000001, giou_loss:   2.36, conf_loss:   0.03, prob_loss:   2.64, total_loss:   5.03\n",
            " Total Loss  tf.Tensor(4.1609144, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  104/125, lr:0.000001, giou_loss:   1.98, conf_loss:   0.11, prob_loss:   2.07, total_loss:   4.16\n",
            " Total Loss  tf.Tensor(4.638797, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  105/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.16, prob_loss:   2.48, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(3.6955633, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  106/125, lr:0.000001, giou_loss:   1.88, conf_loss:   0.09, prob_loss:   1.72, total_loss:   3.70\n",
            " Total Loss  tf.Tensor(4.7743635, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  107/125, lr:0.000001, giou_loss:   2.14, conf_loss:   0.07, prob_loss:   2.56, total_loss:   4.77\n",
            " Total Loss  tf.Tensor(4.292473, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  108/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.10, prob_loss:   2.21, total_loss:   4.29\n",
            " Total Loss  tf.Tensor(4.1044674, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  109/125, lr:0.000001, giou_loss:   1.88, conf_loss:   0.03, prob_loss:   2.20, total_loss:   4.10\n",
            " Total Loss  tf.Tensor(6.0500298, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  110/125, lr:0.000001, giou_loss:   2.58, conf_loss:   0.62, prob_loss:   2.84, total_loss:   6.05\n",
            " Total Loss  tf.Tensor(5.125169, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  111/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.79, prob_loss:   2.23, total_loss:   5.13\n",
            " Total Loss  tf.Tensor(5.400318, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  112/125, lr:0.000001, giou_loss:   2.52, conf_loss:   0.13, prob_loss:   2.75, total_loss:   5.40\n",
            " Total Loss  tf.Tensor(4.5798416, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  113/125, lr:0.000001, giou_loss:   2.13, conf_loss:   0.11, prob_loss:   2.33, total_loss:   4.58\n",
            " Total Loss  tf.Tensor(3.9970737, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  114/125, lr:0.000001, giou_loss:   1.94, conf_loss:   0.02, prob_loss:   2.04, total_loss:   4.00\n",
            " Total Loss  tf.Tensor(5.4128337, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  115/125, lr:0.000001, giou_loss:   2.59, conf_loss:   0.24, prob_loss:   2.58, total_loss:   5.41\n",
            " Total Loss  tf.Tensor(6.1276064, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  116/125, lr:0.000001, giou_loss:   2.74, conf_loss:   0.14, prob_loss:   3.25, total_loss:   6.13\n",
            " Total Loss  tf.Tensor(4.019632, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  117/125, lr:0.000001, giou_loss:   1.80, conf_loss:   0.05, prob_loss:   2.17, total_loss:   4.02\n",
            " Total Loss  tf.Tensor(4.090329, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  118/125, lr:0.000001, giou_loss:   1.89, conf_loss:   0.02, prob_loss:   2.18, total_loss:   4.09\n",
            " Total Loss  tf.Tensor(4.810992, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  119/125, lr:0.000001, giou_loss:   2.26, conf_loss:   0.03, prob_loss:   2.52, total_loss:   4.81\n",
            " Total Loss  tf.Tensor(4.7382393, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  120/125, lr:0.000001, giou_loss:   1.96, conf_loss:   0.25, prob_loss:   2.53, total_loss:   4.74\n",
            " Total Loss  tf.Tensor(4.61963, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  121/125, lr:0.000001, giou_loss:   2.05, conf_loss:   0.13, prob_loss:   2.44, total_loss:   4.62\n",
            " Total Loss  tf.Tensor(4.01457, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  122/125, lr:0.000001, giou_loss:   1.84, conf_loss:   0.15, prob_loss:   2.03, total_loss:   4.01\n",
            " Total Loss  tf.Tensor(4.2454004, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  123/125, lr:0.000001, giou_loss:   1.99, conf_loss:   0.06, prob_loss:   2.19, total_loss:   4.25\n",
            " Total Loss  tf.Tensor(4.6410728, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:  124/125, lr:0.000001, giou_loss:   2.10, conf_loss:   0.06, prob_loss:   2.49, total_loss:   4.64\n",
            " Total Loss  tf.Tensor(4.3689394, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    0/125, lr:0.000001, giou_loss:   2.00, conf_loss:   0.18, prob_loss:   2.20, total_loss:   4.37\n",
            " Total Loss  tf.Tensor(5.3159914, shape=(), dtype=float32)\n",
            " ****  \n",
            "epoch:99 step:    1/125, lr:0.000001, giou_loss:   2.38, conf_loss:   0.14, prob_loss:   2.79, total_loss:   5.32\n",
            "\n",
            "\n",
            "giou_val_loss:   2.18, conf_val_loss:   0.62, prob_val_loss:   5.29, total_val_loss:   8.09\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_MToJJ-zv3FK",
        "colab_type": "text"
      },
      "source": [
        "### Inference"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H96FHkb_AClw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "yolo.save_weights(\"CHECK6/check6\")"
      ],
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rW-APdjCALbx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "b271a900-9069-4fd8-cefc-f3f0f20d859a"
      },
      "source": [
        "!zip -r CHECK6.zip CHECK6/"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  adding: CHECK6/ (stored 0%)\n",
            "  adding: CHECK6/check6.index (deflated 81%)\n",
            "  adding: CHECK6/checkpoint (deflated 41%)\n",
            "  adding: CHECK6/check6.data-00000-of-00001 (deflated 7%)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2vFLqfjVu2_D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "infer = YOLOV3(10, TRAIN=False)\n",
        "tert = np.random.rand(1,416,416,3).astype(np.float32)\n",
        "bert = infer(tert)"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NMn_olSxu23T",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0ad05f62-14d4-4585-e604-52b1d972098b"
      },
      "source": [
        "infer.load_weights(\"CHECK6/check6\")"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x7f8a2e3754e0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2JXCneJZAbS7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!cp CHECK6.zip /content/drive/My\\ Drive/delete/YOLO/"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gbj3qsdcu2bq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import cv2\n",
        "import time\n",
        "import random\n",
        "import colorsys\n",
        "import numpy as np\n",
        "import tensorflow as tf"
      ],
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hG9PsKFIxvZ5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "YOLO_COCO_CLASSES = 'model_data/coco/coco.names'"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XlrbD4_LLs8C",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def draw_bbox(image, bboxes, CLASSES=YOLO_COCO_CLASSES, show_label=True, show_confidence = True, Text_colors=(255,255,0), rectangle_colors='', tracking=False):   \n",
        "    NUM_CLASS = read_class_names(CLASSES)\n",
        "    num_classes = len(NUM_CLASS)\n",
        "    image_h, image_w, _ = image.shape\n",
        "    hsv_tuples = [(1.0 * x / num_classes, 1., 1.) for x in range(num_classes)]\n",
        "    #print(\"hsv_tuples\", hsv_tuples)\n",
        "    colors = list(map(lambda x: colorsys.hsv_to_rgb(*x), hsv_tuples))\n",
        "    colors = list(map(lambda x: (int(x[0] * 255), int(x[1] * 255), int(x[2] * 255)), colors))\n",
        "\n",
        "    random.seed(0)\n",
        "    random.shuffle(colors)\n",
        "    random.seed(None)\n",
        "\n",
        "    for i, bbox in enumerate(bboxes):\n",
        "        coor = np.array(bbox[:4], dtype=np.int32)\n",
        "        score = bbox[4]\n",
        "        class_ind = int(bbox[5])\n",
        "        bbox_color = rectangle_colors if rectangle_colors != '' else colors[class_ind]\n",
        "        bbox_thick = int(0.6 * (image_h + image_w) / 1000)\n",
        "        if bbox_thick < 1: bbox_thick = 1\n",
        "        fontScale = 0.75 * bbox_thick\n",
        "        (x1, y1), (x2, y2) = (coor[0], coor[1]), (coor[2], coor[3])\n",
        "\n",
        "        # put object rectangle\n",
        "        cv2.rectangle(image, (x1, y1), (x2, y2), bbox_color, bbox_thick*2)\n",
        "\n",
        "        if show_label:\n",
        "            # get text label\n",
        "            score_str = \" {:.2f}\".format(score) if show_confidence else \"\"\n",
        "\n",
        "            if tracking: score_str = \" \"+str(score)\n",
        "\n",
        "            label = \"{}\".format(NUM_CLASS[class_ind]) + score_str\n",
        "\n",
        "            # get text size\n",
        "            (text_width, text_height), baseline = cv2.getTextSize(label, cv2.FONT_HERSHEY_COMPLEX_SMALL,\n",
        "                                                                  fontScale, thickness=bbox_thick)\n",
        "            # put filled text rectangle\n",
        "            cv2.rectangle(image, (x1, y1), (x1 + text_width, y1 - text_height - baseline), bbox_color, thickness=cv2.FILLED)\n",
        "\n",
        "            # put text above rectangle\n",
        "            cv2.putText(image, label, (x1, y1-4), cv2.FONT_HERSHEY_COMPLEX_SMALL,\n",
        "                        fontScale, Text_colors, bbox_thick, lineType=cv2.LINE_AA)\n",
        "\n",
        "    return image\n",
        "def bboxes_iou(boxes1, boxes2):\n",
        "    boxes1 = np.array(boxes1)\n",
        "    boxes2 = np.array(boxes2)\n",
        "\n",
        "    boxes1_area = (boxes1[..., 2] - boxes1[..., 0]) * (boxes1[..., 3] - boxes1[..., 1])\n",
        "    boxes2_area = (boxes2[..., 2] - boxes2[..., 0]) * (boxes2[..., 3] - boxes2[..., 1])\n",
        "\n",
        "    left_up       = np.maximum(boxes1[..., :2], boxes2[..., :2])\n",
        "    right_down    = np.minimum(boxes1[..., 2:], boxes2[..., 2:])\n",
        "\n",
        "    inter_section = np.maximum(right_down - left_up, 0.0)\n",
        "    inter_area    = inter_section[..., 0] * inter_section[..., 1]\n",
        "    union_area    = boxes1_area + boxes2_area - inter_area\n",
        "    ious          = np.maximum(1.0 * inter_area / union_area, np.finfo(np.float32).eps)\n",
        "\n",
        "    return ious\n",
        "def nms(bboxes, iou_threshold, sigma=0.3, method='nms'):\n",
        "    \"\"\"\n",
        "    :param bboxes: (xmin, ymin, xmax, ymax, score, class)\n",
        "\n",
        "    Note: soft-nms, https://arxiv.org/pdf/1704.04503.pdf\n",
        "          https://github.com/bharatsingh430/soft-nms\n",
        "    \"\"\"\n",
        "    classes_in_img = list(set(bboxes[:, 5]))\n",
        "    best_bboxes = []\n",
        "\n",
        "    for cls in classes_in_img:\n",
        "        cls_mask = (bboxes[:, 5] == cls)\n",
        "        cls_bboxes = bboxes[cls_mask]\n",
        "        # Process 1: Determine whether the number of bounding boxes is greater than 0 \n",
        "        while len(cls_bboxes) > 0:\n",
        "            # Process 2: Select the bounding box with the highest score according to socre order A\n",
        "            max_ind = np.argmax(cls_bboxes[:, 4])\n",
        "            best_bbox = cls_bboxes[max_ind]\n",
        "            best_bboxes.append(best_bbox)\n",
        "            cls_bboxes = np.concatenate([cls_bboxes[: max_ind], cls_bboxes[max_ind + 1:]])\n",
        "            # Process 3: Calculate this bounding box A and\n",
        "            # Remain all iou of the bounding box and remove those bounding boxes whose iou value is higher than the threshold \n",
        "            iou = bboxes_iou(best_bbox[np.newaxis, :4], cls_bboxes[:, :4])\n",
        "            weight = np.ones((len(iou),), dtype=np.float32)\n",
        "\n",
        "            assert method in ['nms', 'soft-nms']\n",
        "\n",
        "            if method == 'nms':\n",
        "                iou_mask = iou > iou_threshold\n",
        "                weight[iou_mask] = 0.0\n",
        "\n",
        "            if method == 'soft-nms':\n",
        "                weight = np.exp(-(1.0 * iou ** 2 / sigma))\n",
        "\n",
        "            cls_bboxes[:, 4] = cls_bboxes[:, 4] * weight\n",
        "            score_mask = cls_bboxes[:, 4] > 0.\n",
        "            cls_bboxes = cls_bboxes[score_mask]\n",
        "\n",
        "    return best_bboxes\n",
        "def postprocess_boxes(pred_bbox, original_image, input_size, score_threshold):\n",
        "    valid_scale=[0, np.inf]\n",
        "    pred_bbox = np.array(pred_bbox)\n",
        "\n",
        "    pred_xywh = pred_bbox[:, 0:4]\n",
        "    pred_conf = pred_bbox[:, 4]\n",
        "    pred_prob = pred_bbox[:, 5:]\n",
        "\n",
        "    # 1. (x, y, w, h) --> (xmin, ymin, xmax, ymax)\n",
        "    pred_coor = np.concatenate([pred_xywh[:, :2] - pred_xywh[:, 2:] * 0.5,\n",
        "                                pred_xywh[:, :2] + pred_xywh[:, 2:] * 0.5], axis=-1)\n",
        "    # 2. (xmin, ymin, xmax, ymax) -> (xmin_org, ymin_org, xmax_org, ymax_org)\n",
        "    org_h, org_w = original_image.shape[:2]\n",
        "    resize_ratio = min(input_size / org_w, input_size / org_h)\n",
        "\n",
        "    dw = (input_size - resize_ratio * org_w) / 2\n",
        "    dh = (input_size - resize_ratio * org_h) / 2\n",
        "\n",
        "    pred_coor[:, 0::2] = 1.0 * (pred_coor[:, 0::2] - dw) / resize_ratio\n",
        "    pred_coor[:, 1::2] = 1.0 * (pred_coor[:, 1::2] - dh) / resize_ratio\n",
        "\n",
        "    # 3. clip some boxes those are out of range\n",
        "    pred_coor = np.concatenate([np.maximum(pred_coor[:, :2], [0, 0]),\n",
        "                                np.minimum(pred_coor[:, 2:], [org_w - 1, org_h - 1])], axis=-1)\n",
        "    invalid_mask = np.logical_or((pred_coor[:, 0] > pred_coor[:, 2]), (pred_coor[:, 1] > pred_coor[:, 3]))\n",
        "    pred_coor[invalid_mask] = 0\n",
        "\n",
        "    # 4. discard some invalid boxes\n",
        "    bboxes_scale = np.sqrt(np.multiply.reduce(pred_coor[:, 2:4] - pred_coor[:, 0:2], axis=-1))\n",
        "    scale_mask = np.logical_and((valid_scale[0] < bboxes_scale), (bboxes_scale < valid_scale[1]))\n",
        "\n",
        "    # 5. discard boxes with low scores\n",
        "    classes = np.argmax(pred_prob, axis=-1)\n",
        "    scores = pred_conf * pred_prob[np.arange(len(pred_coor)), classes]\n",
        "    score_mask = scores > score_threshold\n",
        "    mask = np.logical_and(scale_mask, score_mask)\n",
        "    coors, scores, classes = pred_coor[mask], scores[mask], classes[mask]\n",
        "\n",
        "    return np.concatenate([coors, scores[:, np.newaxis], classes[:, np.newaxis]], axis=-1)\n",
        "def image_preprocess(image, target_size, gt_boxes=None):\n",
        "    ih, iw    = target_size\n",
        "    h,  w, _  = image.shape\n",
        "\n",
        "    scale = min(iw/w, ih/h)\n",
        "    nw, nh  = int(scale * w), int(scale * h)\n",
        "    image_resized = cv2.resize(image, (nw, nh))\n",
        "\n",
        "    image_paded = np.full(shape=[ih, iw, 3], fill_value=128.0)\n",
        "    dw, dh = (iw - nw) // 2, (ih-nh) // 2\n",
        "    image_paded[dh:nh+dh, dw:nw+dw, :] = image_resized\n",
        "    image_paded = image_paded / 255.\n",
        "\n",
        "    if gt_boxes is None:\n",
        "        return image_paded\n",
        "\n",
        "    else:\n",
        "        gt_boxes[:, [0, 2]] = gt_boxes[:, [0, 2]] * scale + dw\n",
        "        gt_boxes[:, [1, 3]] = gt_boxes[:, [1, 3]] * scale + dh\n",
        "        return image_paded, gt_boxes\n"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MDAraK1Av84v",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def detect_image(YoloV3, image_path, output_path, input_size=416, show=False, CLASSES=YOLO_COCO_CLASSES, score_threshold=0.3, iou_threshold=0.45, rectangle_colors=''):\n",
        "    original_image      = cv2.imread(image_path)\n",
        "    original_image      = cv2.cvtColor(original_image, cv2.COLOR_BGR2RGB)\n",
        "    original_image      = cv2.cvtColor(original_image, cv2.COLOR_BGR2RGB)\n",
        "\n",
        "    image_data = image_preprocess(np.copy(original_image), [input_size, input_size])\n",
        "    image_data = tf.expand_dims(image_data, 0)\n",
        "\n",
        "    pred_bbox = YoloV3.predict(image_data)\n",
        "    pred_bbox = [tf.reshape(x, (-1, tf.shape(x)[-1])) for x in pred_bbox]\n",
        "    pred_bbox = tf.concat(pred_bbox, axis=0)\n",
        "    \n",
        "    bboxes = postprocess_boxes(pred_bbox, original_image, input_size, score_threshold)\n",
        "    print(\" postprocess_boxes\" , len(bboxes))\n",
        "    bboxes = nms(bboxes, iou_threshold, method='nms')\n",
        "    print(\" nms\" , len(bboxes))\n",
        "\n",
        "    image = draw_bbox(original_image, bboxes, CLASSES=CLASSES, rectangle_colors=rectangle_colors)\n",
        "\n",
        "    if output_path != '': cv2.imwrite(output_path, image)\n",
        "    if show:\n",
        "        # Show the image\n",
        "        cv2.imshow(\"predicted image\", image)\n",
        "        # Load and hold the image\n",
        "        cv2.waitKey(0)\n",
        "        # To close the window after the required kill value was provided\n",
        "        cv2.destroyAllWindows()\n",
        "        \n",
        "    return image"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S9HmWEiOv88C",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# image_path   = \"IMAGES/street.jpg\"\n",
        "image_path = 'mnist/mnist_test/000045.jpg'"
      ],
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yzorR0Hev9J8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 173
        },
        "outputId": "819a60a9-4c19-47d1-ee9c-7925f561daec"
      },
      "source": [
        "image = detect_image(infer, image_path, '/content/test1.jpg', input_size=416, CLASSES=TRAIN_CLASSES ,show=False, rectangle_colors=(255,0,0))\n",
        "image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:Layer yolo_v3_1 is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because its dtype defaults to floatx.\n",
            "\n",
            "If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.\n",
            "\n",
            "To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.\n",
            "\n",
            " postprocess_boxes 21\n",
            " nms 2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K96ig-U9yCIi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qJJdnCkKv9Cg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 883
        },
        "outputId": "a461bbd9-e5df-4e09-9ddc-f8aca4f08c26"
      },
      "source": [
        "plt.figure(figsize=(30,15))\n",
        "plt.imshow(image)"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f8a2ccb26a0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 2160x1080 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_qmXtqQIZhaL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OvbbzYeNZhdA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hQ8y7opeZhqR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x511Itt_Zh11",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rU5RoQKSZhxy",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "b82a2f03-4195-4d74-d4e3-ca328d306b08"
      },
      "source": [
        "from evaluate_mAP import get_mAP"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "RuntimeError in tf.config.experimental.list_physical_devices('GPU')\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "O0lgytKNZhjC",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "outputId": "ef1582cb-4d14-4a42-bccc-e6142b4b9444"
      },
      "source": [
        "get_mAP(infer, testset, score_threshold=TEST_SCORE_THRESHOLD, iou_threshold=TEST_IOU_THRESHOLD)"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "calculating mAP45...\n",
            "\n",
            "99.175% = 0 AP  \n",
            "97.340% = 1 AP  \n",
            "96.898% = 2 AP  \n",
            "96.793% = 3 AP  \n",
            "97.179% = 4 AP  \n",
            "95.749% = 5 AP  \n",
            "91.657% = 6 AP  \n",
            "97.533% = 7 AP  \n",
            "99.165% = 8 AP  \n",
            "96.606% = 9 AP  \n",
            "mAP = 96.810%, 14.98 FPS\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "96.80962350978155"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jCon7zDsv813",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}