Minicontest_3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "from google.colab import files\n",
        "files.upload()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 91
        },
        "id": "pEC1V8MAdaF5",
        "outputId": "e81ea085-72eb-4a05-9e72-9e3f3c1854d9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-c781909b-8d37-4752-8c4f-8faa038a4cdd\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-c781909b-8d37-4752-8c4f-8faa038a4cdd\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script>// Copyright 2017 Google LLC\n",
              "//\n",
              "// Licensed under the Apache License, Version 2.0 (the \"License\");\n",
              "// you may not use this file except in compliance with the License.\n",
              "// You may obtain a copy of the License at\n",
              "//\n",
              "//      http://www.apache.org/licenses/LICENSE-2.0\n",
              "//\n",
              "// Unless required by applicable law or agreed to in writing, software\n",
              "// distributed under the License is distributed on an \"AS IS\" BASIS,\n",
              "// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
              "// See the License for the specific language governing permissions and\n",
              "// limitations under the License.\n",
              "\n",
              "/**\n",
              " * @fileoverview Helpers for google.colab Python module.\n",
              " */\n",
              "(function(scope) {\n",
              "function span(text, styleAttributes = {}) {\n",
              "  const element = document.createElement('span');\n",
              "  element.textContent = text;\n",
              "  for (const key of Object.keys(styleAttributes)) {\n",
              "    element.style[key] = styleAttributes[key];\n",
              "  }\n",
              "  return element;\n",
              "}\n",
              "\n",
              "// Max number of bytes which will be uploaded at a time.\n",
              "const MAX_PAYLOAD_SIZE = 100 * 1024;\n",
              "\n",
              "function _uploadFiles(inputId, outputId) {\n",
              "  const steps = uploadFilesStep(inputId, outputId);\n",
              "  const outputElement = document.getElementById(outputId);\n",
              "  // Cache steps on the outputElement to make it available for the next call\n",
              "  // to uploadFilesContinue from Python.\n",
              "  outputElement.steps = steps;\n",
              "\n",
              "  return _uploadFilesContinue(outputId);\n",
              "}\n",
              "\n",
              "// This is roughly an async generator (not supported in the browser yet),\n",
              "// where there are multiple asynchronous steps and the Python side is going\n",
              "// to poll for completion of each step.\n",
              "// This uses a Promise to block the python side on completion of each step,\n",
              "// then passes the result of the previous step as the input to the next step.\n",
              "function _uploadFilesContinue(outputId) {\n",
              "  const outputElement = document.getElementById(outputId);\n",
              "  const steps = outputElement.steps;\n",
              "\n",
              "  const next = steps.next(outputElement.lastPromiseValue);\n",
              "  return Promise.resolve(next.value.promise).then((value) => {\n",
              "    // Cache the last promise value to make it available to the next\n",
              "    // step of the generator.\n",
              "    outputElement.lastPromiseValue = value;\n",
              "    return next.value.response;\n",
              "  });\n",
              "}\n",
              "\n",
              "/**\n",
              " * Generator function which is called between each async step of the upload\n",
              " * process.\n",
              " * @param {string} inputId Element ID of the input file picker element.\n",
              " * @param {string} outputId Element ID of the output display.\n",
              " * @return {!Iterable<!Object>} Iterable of next steps.\n",
              " */\n",
              "function* uploadFilesStep(inputId, outputId) {\n",
              "  const inputElement = document.getElementById(inputId);\n",
              "  inputElement.disabled = false;\n",
              "\n",
              "  const outputElement = document.getElementById(outputId);\n",
              "  outputElement.innerHTML = '';\n",
              "\n",
              "  const pickedPromise = new Promise((resolve) => {\n",
              "    inputElement.addEventListener('change', (e) => {\n",
              "      resolve(e.target.files);\n",
              "    });\n",
              "  });\n",
              "\n",
              "  const cancel = document.createElement('button');\n",
              "  inputElement.parentElement.appendChild(cancel);\n",
              "  cancel.textContent = 'Cancel upload';\n",
              "  const cancelPromise = new Promise((resolve) => {\n",
              "    cancel.onclick = () => {\n",
              "      resolve(null);\n",
              "    };\n",
              "  });\n",
              "\n",
              "  // Wait for the user to pick the files.\n",
              "  const files = yield {\n",
              "    promise: Promise.race([pickedPromise, cancelPromise]),\n",
              "    response: {\n",
              "      action: 'starting',\n",
              "    }\n",
              "  };\n",
              "\n",
              "  cancel.remove();\n",
              "\n",
              "  // Disable the input element since further picks are not allowed.\n",
              "  inputElement.disabled = true;\n",
              "\n",
              "  if (!files) {\n",
              "    return {\n",
              "      response: {\n",
              "        action: 'complete',\n",
              "      }\n",
              "    };\n",
              "  }\n",
              "\n",
              "  for (const file of files) {\n",
              "    const li = document.createElement('li');\n",
              "    li.append(span(file.name, {fontWeight: 'bold'}));\n",
              "    li.append(span(\n",
              "        `(${file.type || 'n/a'}) - ${file.size} bytes, ` +\n",
              "        `last modified: ${\n",
              "            file.lastModifiedDate ? file.lastModifiedDate.toLocaleDateString() :\n",
              "                                    'n/a'} - `));\n",
              "    const percent = span('0% done');\n",
              "    li.appendChild(percent);\n",
              "\n",
              "    outputElement.appendChild(li);\n",
              "\n",
              "    const fileDataPromise = new Promise((resolve) => {\n",
              "      const reader = new FileReader();\n",
              "      reader.onload = (e) => {\n",
              "        resolve(e.target.result);\n",
              "      };\n",
              "      reader.readAsArrayBuffer(file);\n",
              "    });\n",
              "    // Wait for the data to be ready.\n",
              "    let fileData = yield {\n",
              "      promise: fileDataPromise,\n",
              "      response: {\n",
              "        action: 'continue',\n",
              "      }\n",
              "    };\n",
              "\n",
              "    // Use a chunked sending to avoid message size limits. See b/62115660.\n",
              "    let position = 0;\n",
              "    do {\n",
              "      const length = Math.min(fileData.byteLength - position, MAX_PAYLOAD_SIZE);\n",
              "      const chunk = new Uint8Array(fileData, position, length);\n",
              "      position += length;\n",
              "\n",
              "      const base64 = btoa(String.fromCharCode.apply(null, chunk));\n",
              "      yield {\n",
              "        response: {\n",
              "          action: 'append',\n",
              "          file: file.name,\n",
              "          data: base64,\n",
              "        },\n",
              "      };\n",
              "\n",
              "      let percentDone = fileData.byteLength === 0 ?\n",
              "          100 :\n",
              "          Math.round((position / fileData.byteLength) * 100);\n",
              "      percent.textContent = `${percentDone}% done`;\n",
              "\n",
              "    } while (position < fileData.byteLength);\n",
              "  }\n",
              "\n",
              "  // All done.\n",
              "  yield {\n",
              "    response: {\n",
              "      action: 'complete',\n",
              "    }\n",
              "  };\n",
              "}\n",
              "\n",
              "scope.google = scope.google || {};\n",
              "scope.google.colab = scope.google.colab || {};\n",
              "scope.google.colab._files = {\n",
              "  _uploadFiles,\n",
              "  _uploadFilesContinue,\n",
              "};\n",
              "})(self);\n",
              "</script> "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Saving kaggle.json to kaggle.json\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'kaggle.json': b'{\"username\":\"vlb2001\",\"key\":\"d0e9495b8435d2a615da55cc963c2dd5\"}'}"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!rm -r ~/.kaggle\n",
        "!mkdir ~/.kaggle\n",
        "!mv ./kaggle.json ~/.kaggle/\n",
        "!chmod 600 ~/.kaggle/kaggle.json"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2ooxbsuWda37",
        "outputId": "70dce9bb-c67f-4f3a-a6c9-e2f3a9a19a8a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "rm: cannot remove '/root/.kaggle': No such file or directory\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!kaggle competitions download -c uni-na-machine-learning-23-24-ii-sem-contest-3"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-UJq0g-udoCa",
        "outputId": "705ee424-d63f-4cf7-8014-9e0721e7f0b6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading uni-na-machine-learning-23-24-ii-sem-contest-3.zip to /content\n",
            "100% 188M/189M [00:11<00:00, 21.2MB/s]\n",
            "100% 189M/189M [00:11<00:00, 17.6MB/s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import zipfile\n",
        "zip_ref = zipfile.ZipFile('/content/uni-na-machine-learning-23-24-ii-sem-contest-3.zip', 'r')\n",
        "zip_ref.extractall('/content')\n",
        "zip_ref.close()"
      ],
      "metadata": {
        "id": "tS8XOnx0dsZy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GU1CqptU_EOn"
      },
      "source": [
        "#Import di librerie"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R0SMgzYQ__fw"
      },
      "source": [
        "###########################\n",
        "# IMPORT\n",
        "import scipy.io as sio\n",
        "import pandas as pd\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "from torch.optim import lr_scheduler\n",
        "import numpy as np\n",
        "import torchvision\n",
        "from torchvision import datasets, models, transforms\n",
        "import matplotlib.pyplot as plt\n",
        "import time\n",
        "import os\n",
        "import shutil\n",
        "import copy\n",
        "import random\n",
        "from random import shuffle\n",
        "from torch.utils.data import Dataset\n",
        "from PIL import Image\n",
        "import matplotlib.pyplot as plt\n",
        "######################################\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Un primo preprocessing prima dell'elaborazione vera e propria"
      ],
      "metadata": {
        "id": "4Bh1kaNKBq5C"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Definisci il percorso della cartella principale \"train\" e della sottocartella \"train\"\n",
        "train_folder = '/content/train'\n",
        "sub_train_folder = '/content/train/train'\n",
        "test_folder = '/content/test'\n",
        "sub_test_folder = '/content/test/test'\n",
        "\n",
        "\n",
        "#Sposta i file dalla sottocartella alla cartella principale\n",
        "for filename in os.listdir(sub_train_folder):\n",
        "    source = os.path.join(sub_train_folder, filename)\n",
        "    destination = os.path.join(train_folder, filename)\n",
        "    shutil.move(source, destination)\n",
        "\n",
        "for filename in os.listdir(sub_test_folder):\n",
        "    source = os.path.join(sub_test_folder, filename)\n",
        "    destination = os.path.join(test_folder, filename)\n",
        "    shutil.move(source, destination)\n",
        "\n",
        "\n",
        "# Rimuovi la sottocartella vuota \"train\"\n",
        "os.rmdir(sub_train_folder)\n",
        "os.rmdir(sub_test_folder)\n",
        "\n",
        "print(\"File del dataset spostati con successo nella cartella 'train'!\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZeKc_jhtBlM_",
        "outputId": "afb86fab-fd8e-4472-cc06-8031cfcb7c26"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "File del dataset spostati con successo nella cartella 'train'!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pIqUXUmoA541"
      },
      "source": [
        "#Check device\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P2pe5Zh2A4Ui",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b2780ba3-8204-4ff3-803c-7251124fc9c5"
      },
      "source": [
        "print(torch.__version__)\n",
        "print(torch.cuda.is_available())\n",
        "print(torch.cuda.get_device_name(0))\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "print(device)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2.3.0+cu121\n",
            "True\n",
            "Tesla T4\n",
            "cuda:0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YXwdnSlVBrAk"
      },
      "source": [
        "#rendere l'esecuzione deterministica\n",
        "torch.manual_seed(0)\n",
        "np.random.seed(0)\n",
        "torch.backends.cudnn.deterministic = True\n",
        "torch.backends.cudnn.benchmark = False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Classe per lettura dati"
      ],
      "metadata": {
        "id": "nZHueWtR7n0r"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class My_DatasetFolder(Dataset):\n",
        "  def __init__(self, root,  transform, is_valid_file, list_classes):\n",
        "    self.root = root     #path dei dati\n",
        "    self.transform = transform    #lista di trasformazioni\n",
        "    self.is_valid_file = is_valid_file #check per capire quali file leggere\n",
        "    self.list_classes = list_classes   #lista di classi\n",
        "    self.samples = self.__get_samples()\n",
        "\n",
        "\n",
        "  def __len__(self):\n",
        "    return len(self.samples)\n",
        "\n",
        "  # def __get_samples(self):\n",
        "  #   ListFiles=[]\n",
        "  #   for c in self.list_classes:\n",
        "  #     list_folder = os.listdir(self.root + '/' + c)\n",
        "  #     for folder in list_folder:\n",
        "  #       listofFiles = os.listdir(self.root + '/' + c + '/' + folder)\n",
        "  #       for file in listofFiles:\n",
        "  #         if self.is_valid_file(self.root + '/' + c + '/' + folder + '/' + file):\n",
        "  #           ListFiles.append((self.root + '/' + c + '/' + folder + '/' + file, self.list_classes.index(c)))\n",
        "\n",
        "  #   return ListFiles\n",
        "\n",
        "  def __get_samples(self):\n",
        "    ListFiles=[]\n",
        "    for c in self.list_classes:\n",
        "      list_folder = os.listdir(self.root + '/' + c)\n",
        "      for file in list_folder:\n",
        "        if self.is_valid_file(self.root + '/' + c + '/' + file):\n",
        "            ListFiles.append((self.root + '/' + c + '/'+ file, self.list_classes.index(c)))\n",
        "\n",
        "    return ListFiles\n",
        "\n",
        "  def loader_fc(self, path):\n",
        "    img = np.load(path)\n",
        "    return img\n",
        "\n",
        "  def __getitem__(self, index: int):\n",
        "    path, target = self.samples[index]\n",
        "    sample = self.loader_fc(path)\n",
        "    #sample = sample.astype(np.float32)    #qui li converte in float32\n",
        "    if self.transform is not None:\n",
        "      sample = self.transform(sample)\n",
        "    return sample, target\n",
        "      # if target_str is not None:\n",
        "      #     target = 0 if target_str == 'benign' else 1\n",
        "      #     return sample, torch.tensor(target)\n",
        "      # else:\n",
        "      #     return sample, os.path.basename(path)\n",
        "\n",
        "\n",
        "'''\n",
        "un esempio di utilizzo\n",
        "is_valid_file = lambda path: path.endswith('.npy')\n",
        "\n",
        "My_DatasetFolder(root = MyPath,  transform = data_transforms, is_valid_file = is_valid_file, list_classes = ['benign', 'malignant'])\n",
        "'''"
      ],
      "metadata": {
        "id": "T4s9XlV57qe_",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "cdc44e39-ed99-48a6-fb57-9cd92e72fd80"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "\"\\nun esempio di utilizzo\\nis_valid_file = lambda path: path.endswith('.npy')\\n\\nMy_DatasetFolder(root = MyPath,  transform = data_transforms, is_valid_file = is_valid_file, list_classes = ['benign', 'malignant'])\\n\""
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Iperparametri"
      ],
      "metadata": {
        "id": "NTtSe_w1beh8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#iperparametri da scegliere in fase di validation\n",
        "learning_rate = 0.001\n",
        "batch_size = 32\n",
        "set_classes_number = 2\n",
        "num_epoch = 30"
      ],
      "metadata": {
        "id": "BsALaKpzbD5f"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Funzione di Training\n",
        "\n",
        "Funzione utilizzata in fase di training, aggiorna i pesi della rete"
      ],
      "metadata": {
        "id": "z6cMzz6593aj"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7JYaar66_Hec"
      },
      "source": [
        "# scriviamo riga per riga cosa succede durante un addestramento della rete neurael\n",
        "# 1 caricare i dati\n",
        "# 2\n",
        "# 3 confronto con loss\n",
        "# 4 aggiornamento del gradiente\n",
        "# 5\n",
        "\n",
        "# tutto questo per ogni batch, e lo fai per tanti batch fino ad esaurire dati disponibili, cioè hai finito una epoca\n",
        "\n",
        "\n",
        "#find best model on validation test\n",
        "def train_loop_validation(dataloaders, startEpoch, numEpochs, model_conv, criterionCNN, optimizer_conv,\n",
        "                          best_acc, best_loss, best_epoca, outputPath):\n",
        "\n",
        "  for epochs in range(startEpoch, numEpochs + 1):\n",
        "    since = time.time()\n",
        "\n",
        "    modelLoss_train = 0.0\n",
        "    modelAcc_train = 0.0\n",
        "\n",
        "    model_conv.train()\n",
        "\n",
        "    totalSize = 0\n",
        "\n",
        "\n",
        "    #for each batch\n",
        "    for inputs, labels in dataloaders['train']:\n",
        "      inputs = inputs.type(torch.FloatTensor).cuda()\n",
        "      labels = labels.cuda()\n",
        "\n",
        "      # settiamo a zero i gradienti della rete, perchè se no sommi anche quelli precedenti\n",
        "      optimizer_conv.zero_grad()\n",
        "      model_conv.zero_grad()\n",
        "\n",
        "      # output della rete\n",
        "      y = model_conv(inputs)\n",
        "      outp, preds = torch.max(y, 1) # il massimo e indice del massimo\n",
        "\n",
        "      # calcolo la funzione di loss, è un valore solo, la media della loss su un batch\n",
        "      lossCNN = criterionCNN(y, labels) # utilizzo y vera e propria, la softmax è inclusa nella cross entropy\n",
        "\n",
        "      modelLoss_train += lossCNN.item() * inputs.size(0) # media * numero di batch: devo aggregare man mano le somme su tutti i batch\n",
        "      totalSize += inputs.size(0)\n",
        "      modelAcc_train += torch.sum(preds == labels.data).item()\n",
        "\n",
        "      lossCNN.backward()  # pred = f(x)   -> loss = L(f(x), l_true)\n",
        "      optimizer_conv.step()\n",
        "\n",
        "\n",
        "    modelLoss_epoch_train = modelLoss_train/totalSize # divido somma delle varie loss per la size totale del train, allora ho loss su tutto il train\n",
        "    modelAcc_epoch_train  = modelAcc_train/totalSize\n",
        "\n",
        "    #salvataggio dei pesi ad ogni iterazione -> nel caso si blocchi e vogliamo riprendere il train\n",
        "    torch.save(model_conv.state_dict(), outputPath + 'train_weights.pth')\n",
        "\n",
        "    model_conv.eval()\n",
        "    totalSize_val = 0\n",
        "    modelLoss_val = 0.0\n",
        "    modelAcc_val = 0.0\n",
        "\n",
        "    # con il validation sono le stesse cose, tranne per l'aggiornamento dei pesi, perchè dobbiamo solo valutare\n",
        "    for inputs, labels in dataloaders['val']:\n",
        "      inputs = inputs.type(torch.FloatTensor).cuda()\n",
        "      labels = labels.cuda()\n",
        "\n",
        "      y = model_conv(inputs)\n",
        "      outp, preds = torch.max(y, 1)\n",
        "      lossCNN = criterionCNN(y, labels)\n",
        "\n",
        "      modelLoss_val += lossCNN.item() * inputs.size(0)\n",
        "      totalSize_val += inputs.size(0)\n",
        "      modelAcc_val += torch.sum(preds == labels.data).item()\n",
        "\n",
        "\n",
        "    modelLoss_epoch_val=modelLoss_val/totalSize_val\n",
        "    modelAcc_epoch_val = modelAcc_val/totalSize_val\n",
        "    time_elapsed = time.time()-since\n",
        "\n",
        "    #print(time_elapsed)\n",
        "    print('[Epoch %d][TRAIN on %d [Loss: %.4f  ACC: %.4f]][VAL on %d [Loss: %.4f  ACC: %.4f]][TIME: %.0f m %.0f s]'\n",
        "          %(epochs, totalSize, modelLoss_epoch_train, modelAcc_epoch_train, totalSize_val, modelLoss_epoch_val,\n",
        "            modelAcc_epoch_val, time_elapsed // 60, time_elapsed % 60))\n",
        "\n",
        "    # questo è un early stopping se mai dovesse risalire la loss del validation\n",
        "    # così facendo non solo stai salvando i pesi, ma tanto sul validation non stai valutando il modello\n",
        "    # ti serve solo per salvare l'epoca dopo la quale non migliora, cioè salvarti iperparametro che utilizzerai per il test\n",
        "\n",
        "    if epochs == 1 or modelLoss_epoch_val < best_loss:\n",
        "    #if (modelAcc_epoch_val > best_acc) or (modelAcc_epoch_val == best_acc and modelLoss_epoch_val < best_loss) :\n",
        "      print('     .... Saving best weights ....')\n",
        "      best_acc = modelAcc_epoch_val\n",
        "      best_loss = modelLoss_epoch_val\n",
        "      best_epoca = epochs\n",
        "      #salvataggio dei migliori pesi sul validation\n",
        "      torch.save(model_conv.state_dict(), outputPath + 'best_model_weights.pth')\n",
        "\n",
        "    with open(outputPath + 'lossTrain.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelLoss_epoch_train) +'\\n')\n",
        "\n",
        "    with open(outputPath + 'AccTrain.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelAcc_epoch_train)+'\\n')\n",
        "\n",
        "    with open(outputPath + 'lossVal.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelLoss_epoch_val)+'\\n')\n",
        "\n",
        "    with open(outputPath + 'AccVal.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelAcc_epoch_val)+'\\n')\n",
        "\n",
        "    sio.savemat(outputPath + 'check_point.mat', {'best_acc': best_acc,\n",
        "                                                 'best_loss': best_loss,\n",
        "                                                 'best_epoca': best_epoca,\n",
        "                                                 'last_epoch': epochs})"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Conteggio di quante sono ogni acquisizione nelle colture\n",
        "\n",
        "import os\n",
        "\n",
        "train_path = \"/content/train\"  # Sostituisci con il percorso corretto\n",
        "\n",
        "total_images = 0\n",
        "total_benign = 0\n",
        "total_malignant = 0\n",
        "temp_len = 0\n",
        "len_benign_image = []\n",
        "len_malignant_image = []\n",
        "i = 0\n",
        "\n",
        "for class_folder in sorted([\"benign\", \"malignant\"]):  # Itera attraverso le classi\n",
        "    class_path = os.path.join(train_path, class_folder)\n",
        "    for acquisition_folder in sorted(os.listdir(class_path)):\n",
        "        acquisition_path = os.path.join(class_path, acquisition_folder)\n",
        "        if class_folder == \"benign\" and i != 0:\n",
        "            len_benign_image.append(temp_len)\n",
        "            temp_len = 0\n",
        "        if class_folder == \"malignant\" and i != 0:\n",
        "            len_malignant_image.append(temp_len)\n",
        "            temp_len = 0\n",
        "        i += 1\n",
        "        if os.path.isdir(acquisition_path):\n",
        "            for image_file in sorted(os.listdir(acquisition_path)):\n",
        "                if image_file.endswith(\".npy\"):\n",
        "                    temp_len += 1\n",
        "                    if class_folder == \"benign\":\n",
        "                        total_benign += 1\n",
        "                    else:\n",
        "                        total_malignant += 1\n",
        "                    total_images += 1\n",
        "\n",
        "# Append the last temp_len of malignant to len_malignant_image\n",
        "if i != 0:\n",
        "    len_malignant_image.append(temp_len)\n",
        "else:\n",
        "  total_malignant += 1\n",
        "total_images += 1\n",
        "len_benign_image.append(len_malignant_image[0])\n",
        "len_malignant_image = len_malignant_image[1:]\n",
        "\n",
        "print(f\"Numero totale di immagini nel training set: {total_images}\")\n",
        "\n",
        "print(f\"Numero totale di immagini benigne: {total_benign}\")\n",
        "print(f\"Numero totale di immagini maligne: {total_malignant}\")\n",
        "print(f\"Lista immagini benigne: {len_benign_image}\")\n",
        "print(f\"Lista immagini maligne: {len_malignant_image}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3N3mZxsjbq8m",
        "outputId": "7b2e1386-8e6e-469c-99d5-18eb116c38d5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Numero totale di immagini nel training set: 4897\n",
            "Numero totale di immagini benigne: 1277\n",
            "Numero totale di immagini maligne: 3619\n",
            "Lista immagini benigne: [50, 62, 29, 28, 35, 28, 19, 40, 30, 29, 29, 23, 23, 34, 37, 62, 65, 42, 58, 57, 27, 34, 21, 31, 27, 34, 20, 27, 18, 27, 26, 15, 24, 25, 16, 19, 25, 21, 35, 25]\n",
            "Lista immagini maligne: [110, 86, 115, 87, 80, 74, 73, 87, 83, 75, 93, 83, 65, 75, 83, 104, 83, 109, 71, 88, 97, 98, 71, 109, 94, 100, 92, 86, 90, 79, 77, 85, 115, 90, 87, 101, 97, 94, 105, 128]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Creazione Cartelle Locali"
      ],
      "metadata": {
        "id": "UCBsuqjboB7D"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "\n",
        "nomi_cartelle = [\"train_local/benign\",\"train_local/malignant\",\"val_local/benign\", \"val_local/malignant\", \"test_local/benign\",\"test_local/malignant\"]\n",
        "\n",
        "for nome in nomi_cartelle:\n",
        "    # Costruisci il percorso completo della cartella\n",
        "    percorso_cartella = os.path.join('/content', nome)\n",
        "\n",
        "    if not os.path.exists(percorso_cartella):\n",
        "        os.makedirs(percorso_cartella)\n",
        "        print(f\"Cartella '{percorso_cartella}' creata con successo.\")\n",
        "    else:\n",
        "        print(f\"Cartella '{percorso_cartella}' già esistente.\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nenz1sBbk7B7",
        "outputId": "adb68435-797d-4514-dd83-21e096d53f33"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cartella '/content/train_local/benign' già esistente.\n",
            "Cartella '/content/train_local/malignant' già esistente.\n",
            "Cartella '/content/val_local/benign' già esistente.\n",
            "Cartella '/content/val_local/malignant' già esistente.\n",
            "Cartella '/content/test_local/benign' già esistente.\n",
            "Cartella '/content/test_local/malignant' già esistente.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Eliminare cartelle"
      ],
      "metadata": {
        "id": "pF4I9fRQoFXY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# import shutil\n",
        "\n",
        "# # Definisci il nome della cartella che vuoi eliminare\n",
        "# nome_cartella = \"/content/train_local/malignant\"\n",
        "\n",
        "# # Elimina la cartella se esiste\n",
        "# if os.path.exists(nome_cartella):\n",
        "#     shutil.rmtree(nome_cartella)\n",
        "#     print(f\"Cartella '{nome_cartella}' eliminata con successo.\")\n",
        "# else:\n",
        "#     print(f\"Cartella '{nome_cartella}' non trovata.\")"
      ],
      "metadata": {
        "id": "ZXTBhPX5oKCM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Percentuale di benign e malign sull'intero dataset\n",
        "percent_benign_tot = total_benign/total_images\n",
        "percent_malign_tot = 1 - percent_benign_tot\n",
        "\n",
        "print(percent_benign_tot)\n",
        "print(percent_malign_tot)\n",
        "\n",
        "# Percentuale di training, val e test set su intero training set locale\n",
        "len_train_local = total_images*70/100\n",
        "len_val_local = total_images*15/100\n",
        "len_test_local = len_val_local\n",
        "\n",
        "print(len_train_local)\n",
        "print(len_val_local)\n",
        "print(len_test_local)\n",
        "\n",
        "mal_treshold_train_local = len_train_local * percent_malign_tot\n",
        "mal_treshold_val_local = len_val_local * percent_malign_tot\n",
        "mal_treshold_test_local = len_val_local\n",
        "\n",
        "ben_treshold_train_local = len_train_local * percent_benign_tot\n",
        "ben_treshold_val_local = len_val_local * percent_benign_tot\n",
        "mal_treshold_test_local = len_val_local\n",
        "\n",
        "print(mal_treshold_train_local)\n",
        "print(ben_treshold_train_local)\n",
        "\n",
        "# Calcolo del numero di immagini maligne e benigne per training, val e test set\n",
        "num_malign_train = 0\n",
        "num_benign_train = 0\n",
        "num_malign_val = 0\n",
        "num_benign_val = 0\n",
        "num_malign_test = 0\n",
        "num_benign_test = 0\n",
        "\n",
        "import os\n",
        "import shutil\n",
        "\n",
        "# Define the paths for train, val, and test folders\n",
        "train_folder = \"/content/train_local\"\n",
        "val_folder = \"/content/val_local\"\n",
        "test_folder = \"/content/test_local\"\n",
        "\n",
        "# Create the train, val, and test folders if they don't exist\n",
        "os.makedirs(train_folder, exist_ok=True)\n",
        "os.makedirs(val_folder, exist_ok=True)\n",
        "os.makedirs(test_folder, exist_ok=True)\n",
        "\n",
        "# Move the images to the train, val, and test folders based on the calculated thresholds\n",
        "for class_folder in sorted([\"benign\", \"malignant\"]):\n",
        "    class_path = os.path.join(train_path, class_folder)\n",
        "    for acquisition_folder in sorted(os.listdir(class_path)):\n",
        "        acquisition_path = os.path.join(class_path, acquisition_folder)\n",
        "        if os.path.isdir(acquisition_path):\n",
        "            for image_file in sorted(os.listdir(acquisition_path)):\n",
        "                if image_file.endswith(\".npy\"):\n",
        "                    if class_folder == \"malignant\":\n",
        "                        if num_malign_train < mal_treshold_train_local:\n",
        "                            destination_folder = os.path.join(train_folder, class_folder)\n",
        "                            num_malign_train += 1\n",
        "                        elif num_malign_val < mal_treshold_val_local:\n",
        "                            destination_folder = os.path.join(val_folder, class_folder)\n",
        "                            num_malign_val += 1\n",
        "                        else:\n",
        "                            destination_folder = os.path.join(test_folder, class_folder)\n",
        "                            num_malign_test += 1\n",
        "                    else:  # class_folder == \"benign\"\n",
        "                        if num_benign_train < ben_treshold_train_local:\n",
        "                            destination_folder = os.path.join(train_folder, class_folder)\n",
        "                            num_benign_train += 1\n",
        "                        elif num_benign_val < ben_treshold_val_local:\n",
        "                            destination_folder = os.path.join(val_folder, class_folder)\n",
        "                            num_benign_val += 1\n",
        "                        else:\n",
        "                            destination_folder = os.path.join(test_folder, class_folder)\n",
        "                            num_benign_test += 1\n",
        "\n",
        "                    # Create the destination folder if it doesn't exist\n",
        "                    os.makedirs(destination_folder, exist_ok=True)\n",
        "\n",
        "                    source_path = os.path.join(acquisition_path, image_file)\n",
        "                    destination_path = os.path.join(destination_folder, image_file)\n",
        "                    shutil.copyfile(source_path, destination_path)\n",
        "\n",
        "\n",
        "# Print the number of images in each set\n",
        "print(f\"Number of malignant images in train set: {num_malign_train}\")\n",
        "print(f\"Number of benign images in train set: {num_benign_train}\")\n",
        "print(f\"Number of malignant images in val set: {num_malign_val}\")\n",
        "print(f\"Number of benign images in val set: {num_benign_val}\")\n",
        "print(f\"Number of malignant images in test set: {num_malign_test}\")\n",
        "print(f\"Number of benign images in test set: {num_benign_test}\")\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l5IUVRyQgKBr",
        "outputId": "143bb353-333f-461d-8c56-bc2e72eca1d6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.26077190116397797\n",
            "0.739228098836022\n",
            "3427.9\n",
            "734.55\n",
            "734.55\n",
            "2534.0\n",
            "893.9000000000001\n",
            "Number of malignant images in train set: 2534\n",
            "Number of benign images in train set: 894\n",
            "Number of malignant images in val set: 543\n",
            "Number of benign images in val set: 192\n",
            "Number of malignant images in test set: 542\n",
            "Number of benign images in test set: 191\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Caricamento dei Dati"
      ],
      "metadata": {
        "id": "C_HH3rqm-ty6"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EJ3hEKCPCObR"
      },
      "source": [
        "# varie trasformazioni che vengono applicate alle immagini prima che verranno caricate in memoria\n",
        "# trasformazioni su train e val sono diverse, hanno in comune quella in tensor, ma serve solo per formato giusto ingresso rete\n",
        "# classe normalize prende in ingresso due vettori, uno delle medie e uno delle deviazioni standard, ogni vettore è da 3 perchè abbiamo tre canali\n",
        "# è come se avessi fatto ogni canale a media nulla e deviaz std unitaria, i numeri sono stati ottenuti calcolando le medie e dev std sul training\n",
        "# quindi quando lo si fa nel pratico devi calcolarteli prima a parte\n",
        "\n",
        "data_transforms = {\n",
        "    'train_local': transforms.Compose([ # con augumentation irrobustico il sistema, applico delle variabilità per cercare di introdurre generalità nel modello\n",
        "        transforms.ToTensor(), # converto immagine in tensore\n",
        "        transforms.Resize(224),\n",
        "        #transforms.ToPILImage(),\n",
        "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]) # qusti sono quelli ottenuti per dataset iniziale che noi non abbiamo, medie e dev di imagenet (il dataset originale)\n",
        "    ]),\n",
        "    'val_local': transforms.Compose([ # facciamo una cosa controllata sul val e non randomica perchè bisogna fare una prima valutazione sul validation\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Resize(224),\n",
        "        #transforms.ToPILImage(),\n",
        "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
        "    ]),\n",
        "    'test_local': transforms.Compose([ # facciamo una cosa controllata sul val e non randomica perchè bisogna fare una prima valutazione sul validation\n",
        "        transforms.ToTensor(),\n",
        "        transforms.Resize(224),\n",
        "        #transforms.ToPILImage(),\n",
        "        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n",
        "    ]),\n",
        "}\n",
        "\n",
        "# image folder si ricorda dove stanno le immagini senza caricarle in memoria\n",
        "# gestiamo in maniera efficiente i nostri dati, in particolare i percorrisi delle immagini\n",
        "# in attributo samples ci sono tutti i path e le classi associate\n",
        "# in attributo classes ci sono il vettore delle classi, in ordine di indice come sono mostrate\n",
        "# per accedere all'immagine vera e propria la si ha con funzione getitem, così la carichi proprio\n",
        "# però prima che prendiamo un immagine in mem, image folder la converte con le trasformazioini di sopra\n",
        "# ogni immagine in pytorch deve essere un tensore per poter essere elaborata\n",
        "# 1d = numero canali, 2d e 3d grandezza immagine\n",
        "\n",
        "content_dir = '/content'\n",
        "\n",
        "is_valid_file = lambda path: path.endswith('.npy')\n",
        "my_dataset = {x: My_DatasetFolder(os.path.join(content_dir, x),\n",
        "                                          data_transforms[x],is_valid_file=is_valid_file,list_classes=['benign','malignant'])\n",
        "                  for x in ['train_local','val_local','test_local']}\n",
        "\n",
        "\n",
        "\n",
        "# # Il dataloader è la classe responsabile per caricare i dati in memoria\n",
        "\n",
        "dataloaders = {x: torch.utils.data.DataLoader(my_dataset[x], batch_size=batch_size,\n",
        "                                             shuffle=True, num_workers=4)\n",
        "              for x in ['train_local','val_local','test_local']}\n",
        "dataset_sizes = {x: len(my_dataset[x]) for x in ['train_local','val_local','test_local']}\n",
        "class_names = my_dataset['train_local'].list_classes\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "\n",
        "# print(class_names)\n",
        "# print(image_datasets['val'].classes)\n",
        "\n",
        "# # Quando sono stati caricati i dati in mem lo si è fatto in due cartelle separate, una per training e una per validation\n",
        "# # perciò sono stati fatti due imagefolder\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "my_dataset['test_local'].samples\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "collapsed": true,
        "id": "5wPdBUab1Qcc",
        "outputId": "0f6a171d-a223-48b7-8f63-3dd43ec1322b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_22.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_19.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_20.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_20.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_22.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_33.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_28.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_34.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_26.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_21.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_22.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_32.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_21.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_24.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_27.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_22.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_19.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_19.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_21.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_23.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_24.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_19.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_20.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_21.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_20.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_25.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_12.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_23.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_19.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_22.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_30.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_23.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-32_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_20.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_29.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_8.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_23.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_24.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_20.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_31.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_3.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_21.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_6.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_16.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_4.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_0.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_13.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_14.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_1.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-37_croppedCell_24.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_2.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-35_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_7.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-38_croppedCell_10.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_17.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_18.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_15.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-40_croppedCell_23.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-39_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_9.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-36_croppedCell_5.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-34_croppedCell_19.npy', 0),\n",
              " ('/content/test_local/benign/mcf10aCdExposed-33_croppedCell_11.npy', 0),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_75.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_51.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_71.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_28.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_84.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_96.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_0.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_36.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_85.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_3.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_46.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_88.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_17.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_56.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_70.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_90.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_74.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_74.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_4.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_42.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_25.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_8.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_13.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_68.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_56.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_44.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_19.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_39.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_43.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_35.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_119.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_55.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_66.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_103.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_82.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_80.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_27.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_8.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_77.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_45.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_65.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_15.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_69.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_10.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_97.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_54.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_112.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_88.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_55.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_69.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_60.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_56.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_10.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_23.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_53.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_81.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_66.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_10.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_53.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_35.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_84.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_17.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_60.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_104.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_74.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_72.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_105.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_83.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_0.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_84.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_76.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_45.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_82.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_36.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_14.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_59.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_11.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_51.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_79.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_1.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_102.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_96.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_26.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_92.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_62.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_93.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_76.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_98.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_115.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_30.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_85.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_2.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_71.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_20.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_0.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_73.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_69.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_20.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_4.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_12.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_96.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_41.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_30.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_51.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_8.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_91.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_21.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_47.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_80.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_48.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_65.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_75.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_54.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_83.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_19.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_81.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_31.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_46.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_15.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_6.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_123.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_93.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_26.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_48.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_39.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_12.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_9.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_75.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_79.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_12.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_9.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_80.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_3.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_11.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_84.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_93.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_64.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_14.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_16.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_67.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_20.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_30.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_59.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_1.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_87.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_79.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_81.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_31.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_90.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_2.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_17.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_33.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_24.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_40.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_13.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_47.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_28.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_18.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_74.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_51.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_38.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_39.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_18.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_47.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_69.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_15.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_18.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_50.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_35.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_63.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_75.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_44.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_75.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_28.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_91.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_95.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_68.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_19.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_22.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_59.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_120.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_16.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_26.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_71.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_35.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_25.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_52.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_42.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_25.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_63.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_61.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_7.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_59.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_64.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_98.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_59.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_16.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_41.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_67.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_88.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_37.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_17.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_63.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_33.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_44.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_36.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_86.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_50.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_41.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_46.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_11.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_44.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_89.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_81.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_33.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_14.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_82.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_34.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_62.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_102.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_90.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_66.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_54.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_16.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_84.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_9.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_2.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_13.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_71.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_21.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_40.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_19.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_50.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_87.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_53.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_86.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_24.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_73.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_99.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_67.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_0.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_111.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_6.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_79.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_98.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_92.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_47.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_4.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_19.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_78.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_14.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_80.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_3.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_75.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_62.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_27.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_28.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_95.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_29.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_6.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_0.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_67.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_78.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_61.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_23.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_20.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_39.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_63.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_58.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_26.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_6.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_76.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_9.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_73.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_32.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_38.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_57.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_103.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_18.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_72.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_71.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_1.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_13.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_22.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_61.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_87.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_57.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_70.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_69.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_29.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_34.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_63.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_68.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_60.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_53.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_3.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_85.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_77.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_65.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_79.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_76.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_23.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_53.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_32.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_16.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_80.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_30.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_42.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_64.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_46.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_35.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_15.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_49.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_86.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_38.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_43.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_52.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_88.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_31.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_127.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_76.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_51.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_10.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_22.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_74.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_55.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_45.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_33.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_70.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_118.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_58.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_107.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_49.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_72.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_38.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_77.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_58.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_87.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_58.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_37.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_36.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_77.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_54.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_90.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_17.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_50.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_66.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_24.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_8.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_86.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_97.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_70.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_95.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_8.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_7.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_40.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_37.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_82.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_24.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_73.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_86.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_49.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_89.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_66.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_7.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_23.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_15.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_92.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_85.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_42.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_73.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_43.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_27.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_88.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_45.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_93.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_40.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_38.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_57.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_32.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_49.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_37.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_13.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_43.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_5.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_57.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_68.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_9.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_44.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_113.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_90.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_85.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_114.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_4.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_25.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_78.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_5.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_64.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_64.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_58.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_29.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_78.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_33.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_32.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_47.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_76.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_116.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_43.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_37.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_100.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_81.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_62.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_29.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_28.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_89.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_93.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_22.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_83.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_41.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_100.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_23.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_22.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_11.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_5.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_48.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_125.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_121.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_36.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_104.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_126.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_92.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_110.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_60.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_96.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_54.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_83.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_82.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_25.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_78.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_85.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_124.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_5.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_55.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_72.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_48.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_78.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_42.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_72.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_27.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_55.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_94.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_1.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_60.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_97.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_95.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_24.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_29.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_34.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_79.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_9.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_49.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_73.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_117.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_100.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_21.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_99.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_7.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_77.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_2.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_31.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_94.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_20.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_61.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_3.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_34.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_27.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_52.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_82.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_12.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_56.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_109.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_50.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_94.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_41.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_52.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_57.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_83.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_101.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_101.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_91.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_70.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_48.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_10.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_4.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_21.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_21.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_92.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_31.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_65.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_89.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_7.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_11.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_18.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_39.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_86.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_74.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_65.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_122.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_8.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_46.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_72.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_91.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_26.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_61.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_89.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_80.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_81.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_94.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_67.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_77.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_12.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_30.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_52.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_45.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_91.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_34.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_99.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_40.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-35_croppedCell_83.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_14.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_32.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-39_croppedCell_68.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_62.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-36_croppedCell_6.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_87.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_1.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_108.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_84.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_2.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-37_croppedCell_56.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-38_croppedCell_5.npy', 1),\n",
              " ('/content/test_local/malignant/mcf7CdExposed-40_croppedCell_106.npy', 1)]"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Training Function"
      ],
      "metadata": {
        "id": "c77F45qc5PG0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def train_loop_validation(dataloaders, startEpoch, numEpochs, model_conv, criterionCNN, optimizer_conv,\n",
        "                          best_acc, best_loss, best_epoca, outputPath):\n",
        "\n",
        "  for epochs in range(startEpoch, numEpochs + 1):\n",
        "    since = time.time()\n",
        "\n",
        "    modelLoss_train = 0.0\n",
        "    modelAcc_train = 0.0\n",
        "\n",
        "    model_conv.train()\n",
        "\n",
        "    totalSize = 0\n",
        "\n",
        "\n",
        "    #for each batch\n",
        "    for inputs, labels in dataloaders['train_local']:\n",
        "      inputs = inputs.type(torch.FloatTensor).cuda()\n",
        "      labels = labels.cuda()\n",
        "\n",
        "      # settiamo a zero i gradienti della rete, perchè se no sommi anche quelli precedenti\n",
        "      optimizer_conv.zero_grad()\n",
        "      model_conv.zero_grad()\n",
        "\n",
        "      # output della rete\n",
        "      y = model_conv(inputs)\n",
        "      outp, preds = torch.max(y, 1) # il massimo e indice del massimo\n",
        "\n",
        "      # calcolo la funzione di loss, è un valore solo, la media della loss su un batch\n",
        "      lossCNN = criterionCNN(y, labels) # utilizzo y vera e propria, la softmax è inclusa nella cross entropy\n",
        "\n",
        "      modelLoss_train += lossCNN.item() * inputs.size(0) # media * numero di batch: devo aggregare man mano le somme su tutti i batch\n",
        "      totalSize += inputs.size(0)\n",
        "      modelAcc_train += torch.sum(preds == labels.data).item()\n",
        "\n",
        "      lossCNN.backward()  # pred = f(x)   -> loss = L(f(x), l_true)\n",
        "      optimizer_conv.step()\n",
        "\n",
        "\n",
        "    modelLoss_epoch_train = modelLoss_train/totalSize # divido somma delle varie loss per la size totale del train, allora ho loss su tutto il train\n",
        "    modelAcc_epoch_train  = modelAcc_train/totalSize\n",
        "\n",
        "    #salvataggio dei pesi ad ogni iterazione -> nel caso si blocchi e vogliamo riprendere il train\n",
        "    torch.save(model_conv.state_dict(), outputPath + 'train_weights.pth')\n",
        "\n",
        "    model_conv.eval()\n",
        "    totalSize_val = 0\n",
        "    modelLoss_val = 0.0\n",
        "    modelAcc_val = 0.0\n",
        "\n",
        "    # con il validation sono le stesse cose, tranne per l'aggiornamento dei pesi, perchè dobbiamo solo valutare\n",
        "    for inputs, labels in dataloaders['val_local']:\n",
        "      inputs = inputs.type(torch.FloatTensor).cuda()\n",
        "      labels = labels.cuda()\n",
        "\n",
        "      y = model_conv(inputs)\n",
        "      outp, preds = torch.max(y, 1)\n",
        "      lossCNN = criterionCNN(y, labels)\n",
        "\n",
        "      modelLoss_val += lossCNN.item() * inputs.size(0)\n",
        "      totalSize_val += inputs.size(0)\n",
        "      modelAcc_val += torch.sum(preds == labels.data).item()\n",
        "\n",
        "\n",
        "    modelLoss_epoch_val=modelLoss_val/totalSize_val\n",
        "    modelAcc_epoch_val = modelAcc_val/totalSize_val\n",
        "    time_elapsed = time.time()-since\n",
        "\n",
        "    #print(time_elapsed)\n",
        "    print('[Epoch %d][TRAIN on %d [Loss: %.4f  ACC: %.4f]][VAL on %d [Loss: %.4f  ACC: %.4f]][TIME: %.0f m %.0f s]'\n",
        "          %(epochs, totalSize, modelLoss_epoch_train, modelAcc_epoch_train, totalSize_val, modelLoss_epoch_val,\n",
        "            modelAcc_epoch_val, time_elapsed // 60, time_elapsed % 60))\n",
        "\n",
        "    # questo è un early stopping se mai dovesse risalire la loss del validation\n",
        "    # così facendo non solo stai salvando i pesi, ma tanto sul validation non stai valutando il modello\n",
        "    # ti serve solo per salvare l'epoca dopo la quale non migliora, cioè salvarti iperparametro che utilizzerai per il test\n",
        "\n",
        "    if epochs == 1 or modelLoss_epoch_val < best_loss:\n",
        "    #if (modelAcc_epoch_val > best_acc) or (modelAcc_epoch_val == best_acc and modelLoss_epoch_val < best_loss) :\n",
        "      print('     .... Saving best weights ....')\n",
        "      best_acc = modelAcc_epoch_val\n",
        "      best_loss = modelLoss_epoch_val\n",
        "      best_epoca = epochs\n",
        "      #salvataggio dei migliori pesi sul validation\n",
        "      torch.save(model_conv.state_dict(), outputPath + 'best_model_weights.pth')\n",
        "\n",
        "    with open(outputPath + 'lossTrain.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelLoss_epoch_train) +'\\n')\n",
        "\n",
        "    with open(outputPath + 'AccTrain.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelAcc_epoch_train)+'\\n')\n",
        "\n",
        "    with open(outputPath + 'lossVal.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelLoss_epoch_val)+'\\n')\n",
        "\n",
        "    with open(outputPath + 'AccVal.txt', \"a\") as file_object:\n",
        "      file_object.write(str(modelAcc_epoch_val)+'\\n')\n",
        "\n",
        "    sio.savemat(outputPath + 'check_point.mat', {'best_acc': best_acc,\n",
        "                                                 'best_loss': best_loss,\n",
        "                                                 'best_epoca': best_epoca,\n",
        "                                                 'last_epoch': epochs})"
      ],
      "metadata": {
        "id": "kitY3OKh5Ncd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Scelta del Modello"
      ],
      "metadata": {
        "id": "1wgRFLHM2TyO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "'''\n",
        "Inserire il modello che si vuole considerare\n",
        "'''\n",
        "\n",
        "networkName = 'vgg11'\n",
        "WeightPath = '/content/gdrive/My Drive/Colab Notebooks/' + networkName + '/'\n",
        "\n",
        "try:\n",
        "  os.makedirs(WeightPath)\n",
        "except:\n",
        "  pass\n",
        "\n",
        "'''\n",
        "model_ft = models.resnet18(pretrained=True)\n",
        "num_ftrs = model_ft.fc.in_features\n",
        "# Here the size of each output sample is set to 2.\n",
        "# Alternatively, it can be generalized to nn.Linear(num_ftrs, len(class_names)).\n",
        "model_ft.fc = nn.Linear(num_ftrs, 2)\n",
        "\n",
        "for param in model_ft.parameters():\n",
        "  param.requires_grad = True\n",
        "\n",
        "'''\n",
        "\n",
        "model = torchvision.models.vgg11(pretrained=True)\n",
        "# remove last fully-connected layer\n",
        "\n",
        "\n",
        "# change last fully connected layer\n",
        "model.classifier[6] = nn.Linear(4096, set_classes_number)\n",
        "\n",
        "\n",
        "# set requires_grad=false\n",
        "for param in model.parameters():\n",
        "  param.requires_grad = False # tutta la rete si sblocca\n",
        "\n",
        "# for layer in model.features[18:]:\n",
        "#   for param in layer.parameters():\n",
        "#       param.requires_grad = True\n",
        "\n",
        "# train fully connected layers\n",
        "for param in model.classifier.parameters():\n",
        "  param.requires_grad = True # sblocchiamo i layer della parte di classificazione, si aggiornano durante l'addestramento\n",
        "\n",
        "print(model.parameters)\n",
        "\n",
        "\n",
        "# Ora vogliamo addestrare, per prima cosa bisogna scegliere la funzione di loss\n",
        "# per le classi va per la maggiore la cross entropy/ binary cross entropy\n",
        "# ci serve anche un ottimizzatore, cioè quello che effettivamente fa algoritmo iterativo che fa gli step\n",
        "\n",
        "model_conv = model.cuda()\n",
        "criterion = nn.CrossEntropyLoss()\n",
        "optimizer_conv = optim.Adam(model_conv.parameters(), lr=learning_rate)\n",
        "\n",
        "# -------------------------------------------------------------------------------------  train model\n",
        "startEpoch =1\n",
        "best_acc = 0\n",
        "best_loss=0\n",
        "best_epoca = 0\n",
        "\n",
        "# funzione definita precedentemente per fare il training\n",
        "train_loop_validation(dataloaders, startEpoch, num_epoch, model_conv, criterion, optimizer_conv, best_acc, best_loss, best_epoca, WeightPath)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iONBunO62TOS",
        "outputId": "8e01f36b-382d-4399-b973-33d74057cc9a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<bound method Module.parameters of VGG(\n",
            "  (features): Sequential(\n",
            "    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (1): ReLU(inplace=True)\n",
            "    (2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (3): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (4): ReLU(inplace=True)\n",
            "    (5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (6): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (7): ReLU(inplace=True)\n",
            "    (8): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (9): ReLU(inplace=True)\n",
            "    (10): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (11): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (12): ReLU(inplace=True)\n",
            "    (13): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (14): ReLU(inplace=True)\n",
            "    (15): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (16): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (17): ReLU(inplace=True)\n",
            "    (18): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (19): ReLU(inplace=True)\n",
            "    (20): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "  )\n",
            "  (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))\n",
            "  (classifier): Sequential(\n",
            "    (0): Linear(in_features=25088, out_features=4096, bias=True)\n",
            "    (1): ReLU(inplace=True)\n",
            "    (2): Dropout(p=0.5, inplace=False)\n",
            "    (3): Linear(in_features=4096, out_features=4096, bias=True)\n",
            "    (4): ReLU(inplace=True)\n",
            "    (5): Dropout(p=0.5, inplace=False)\n",
            "    (6): Linear(in_features=4096, out_features=2, bias=True)\n",
            "  )\n",
            ")>\n",
            "[Epoch 1][TRAIN on 3428 [Loss: 0.5317  ACC: 0.7900]][VAL on 735 [Loss: 0.6407  ACC: 0.7456]][TIME: 0 m 23 s]\n",
            "     .... Saving best weights ....\n",
            "[Epoch 2][TRAIN on 3428 [Loss: 0.3280  ACC: 0.8819]][VAL on 735 [Loss: 0.6886  ACC: 0.8490]][TIME: 0 m 26 s]\n",
            "[Epoch 3][TRAIN on 3428 [Loss: 0.2801  ACC: 0.9061]][VAL on 735 [Loss: 0.5738  ACC: 0.8286]][TIME: 0 m 23 s]\n",
            "     .... Saving best weights ....\n",
            "[Epoch 4][TRAIN on 3428 [Loss: 0.1949  ACC: 0.9349]][VAL on 735 [Loss: 0.7419  ACC: 0.8286]][TIME: 0 m 24 s]\n",
            "[Epoch 5][TRAIN on 3428 [Loss: 0.2225  ACC: 0.9393]][VAL on 735 [Loss: 0.9156  ACC: 0.8680]][TIME: 0 m 24 s]\n",
            "[Epoch 6][TRAIN on 3428 [Loss: 0.1432  ACC: 0.9641]][VAL on 735 [Loss: 0.9673  ACC: 0.8599]][TIME: 0 m 23 s]\n",
            "[Epoch 7][TRAIN on 3428 [Loss: 0.1382  ACC: 0.9647]][VAL on 735 [Loss: 0.9503  ACC: 0.8721]][TIME: 0 m 23 s]\n",
            "[Epoch 8][TRAIN on 3428 [Loss: 0.1695  ACC: 0.9606]][VAL on 735 [Loss: 0.7777  ACC: 0.8680]][TIME: 0 m 25 s]\n",
            "[Epoch 9][TRAIN on 3428 [Loss: 0.1215  ACC: 0.9667]][VAL on 735 [Loss: 1.0361  ACC: 0.8721]][TIME: 0 m 23 s]\n",
            "[Epoch 10][TRAIN on 3428 [Loss: 0.0972  ACC: 0.9758]][VAL on 735 [Loss: 1.5798  ACC: 0.8558]][TIME: 0 m 24 s]\n",
            "[Epoch 11][TRAIN on 3428 [Loss: 0.0989  ACC: 0.9749]][VAL on 735 [Loss: 1.6340  ACC: 0.8612]][TIME: 0 m 23 s]\n",
            "[Epoch 12][TRAIN on 3428 [Loss: 0.1400  ACC: 0.9726]][VAL on 735 [Loss: 1.6863  ACC: 0.8748]][TIME: 0 m 25 s]\n",
            "[Epoch 13][TRAIN on 3428 [Loss: 0.1223  ACC: 0.9749]][VAL on 735 [Loss: 1.8413  ACC: 0.8367]][TIME: 0 m 25 s]\n",
            "[Epoch 14][TRAIN on 3428 [Loss: 0.1048  ACC: 0.9790]][VAL on 735 [Loss: 1.3033  ACC: 0.8340]][TIME: 0 m 24 s]\n",
            "[Epoch 15][TRAIN on 3428 [Loss: 0.0584  ACC: 0.9863]][VAL on 735 [Loss: 2.3899  ACC: 0.8639]][TIME: 0 m 25 s]\n",
            "[Epoch 16][TRAIN on 3428 [Loss: 0.1545  ACC: 0.9764]][VAL on 735 [Loss: 2.3935  ACC: 0.8612]][TIME: 0 m 24 s]\n",
            "[Epoch 17][TRAIN on 3428 [Loss: 0.2289  ACC: 0.9711]][VAL on 735 [Loss: 2.9456  ACC: 0.7905]][TIME: 0 m 22 s]\n",
            "[Epoch 18][TRAIN on 3428 [Loss: 0.1607  ACC: 0.9784]][VAL on 735 [Loss: 2.4046  ACC: 0.8776]][TIME: 0 m 24 s]\n",
            "[Epoch 19][TRAIN on 3428 [Loss: 0.0864  ACC: 0.9837]][VAL on 735 [Loss: 1.8661  ACC: 0.8707]][TIME: 0 m 23 s]\n",
            "[Epoch 20][TRAIN on 3428 [Loss: 0.0684  ACC: 0.9877]][VAL on 735 [Loss: 1.8722  ACC: 0.8503]][TIME: 0 m 26 s]\n",
            "[Epoch 21][TRAIN on 3428 [Loss: 0.0592  ACC: 0.9901]][VAL on 735 [Loss: 2.4281  ACC: 0.8680]][TIME: 0 m 24 s]\n",
            "[Epoch 22][TRAIN on 3428 [Loss: 0.0770  ACC: 0.9869]][VAL on 735 [Loss: 1.9290  ACC: 0.8680]][TIME: 0 m 22 s]\n",
            "[Epoch 23][TRAIN on 3428 [Loss: 0.0514  ACC: 0.9895]][VAL on 735 [Loss: 3.1373  ACC: 0.8667]][TIME: 0 m 22 s]\n",
            "[Epoch 24][TRAIN on 3428 [Loss: 0.0543  ACC: 0.9901]][VAL on 735 [Loss: 2.1417  ACC: 0.8531]][TIME: 0 m 26 s]\n",
            "[Epoch 25][TRAIN on 3428 [Loss: 0.0557  ACC: 0.9895]][VAL on 735 [Loss: 3.9007  ACC: 0.8585]][TIME: 0 m 23 s]\n",
            "[Epoch 26][TRAIN on 3428 [Loss: 0.1275  ACC: 0.9842]][VAL on 735 [Loss: 4.3611  ACC: 0.8408]][TIME: 0 m 24 s]\n",
            "[Epoch 27][TRAIN on 3428 [Loss: 0.4997  ACC: 0.9644]][VAL on 735 [Loss: 2.7958  ACC: 0.8422]][TIME: 0 m 24 s]\n",
            "[Epoch 28][TRAIN on 3428 [Loss: 0.2719  ACC: 0.9758]][VAL on 735 [Loss: 3.1737  ACC: 0.8571]][TIME: 0 m 24 s]\n",
            "[Epoch 29][TRAIN on 3428 [Loss: 0.1439  ACC: 0.9863]][VAL on 735 [Loss: 4.5456  ACC: 0.8612]][TIME: 0 m 25 s]\n",
            "[Epoch 30][TRAIN on 3428 [Loss: 0.3072  ACC: 0.9799]][VAL on 735 [Loss: 4.4522  ACC: 0.8612]][TIME: 0 m 22 s]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Feature Extraction"
      ],
      "metadata": {
        "id": "ZrnbbxDD1xAg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "outfeatures_train =[]\n",
        "\n",
        "data_transforms = transforms.Compose([\n",
        "                                      transforms.Resize(224),\n",
        "                                      transforms.ToTensor(),\n",
        "                                         #questo è solo un esempio!\n",
        "                                      ])\n",
        "\n",
        "'''\n",
        "Inserire il modello che si vuole considerare\n",
        "'''\n",
        "model = torchvision.models.vgg11(pretrained=True)\n",
        "# remove last fully-connected layer\n",
        "new_classifier = nn.Sequential(*list(model.classifier.children())[:-1])\n",
        "model.classifier = new_classifier\n",
        "model = model.cuda()\n",
        "model.eval()\n",
        "\n",
        "class_list_train = []\n",
        "class_list_test = []\n",
        "classes = ['benign', 'malignant']\n",
        "data_dir = '/content'   #path dei dati\n",
        "\n",
        "for c in classes:\n",
        "  folder = os.listdir(data_dir + '/train/' + c)\n",
        "  for f in folder:\n",
        "    listImage = os.listdir(data_dir + '/train/' + c + '/' + f)\n",
        "    for input in listImage:\n",
        "      img = np.load(data_dir + '/train/' + c + '/' + f + '/' + input)\n",
        "      img_pil = Image.fromarray((img * 255).astype(np.uint8))\n",
        "      img = data_transforms(img_pil).type(torch.FloatTensor).to(device)\n",
        "      img = img.unsqueeze_(0)\n",
        "      outputs = model(img)\n",
        "\n",
        "      outfeatures_train.append((np.squeeze(outputs.cpu().detach()).numpy()))\n",
        "      class_list_train.append(classes.index(c))\n",
        "\n",
        "# Dividiamo il training test in training, validation e test locale\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BBaIr79p9szU",
        "outputId": "bca6cd56-25c5-4409-892a-c156ca459b7a"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n",
            "  warnings.warn(\n",
            "/usr/local/lib/python3.10/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=VGG11_Weights.IMAGENET1K_V1`. You can also use `weights=VGG11_Weights.DEFAULT` to get the most up-to-date weights.\n",
            "  warnings.warn(msg)\n",
            "Downloading: \"https://download.pytorch.org/models/vgg11-8a719046.pth\" to /root/.cache/torch/hub/checkpoints/vgg11-8a719046.pth\n",
            "100%|██████████| 507M/507M [00:06<00:00, 80.5MB/s]\n"
          ]
        }
      ]
    }
  ]
}