-
Notifications
You must be signed in to change notification settings - Fork 26
/
Copy path1_test_dcn.py
269 lines (218 loc) · 12.6 KB
/
1_test_dcn.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
#! /usr/bin/env python
# coding=utf-8
# ================================================================
#
# Author : miemie2013
# Created date: 2020-10-15 14:50:03
# Description : pytorch_ppyolo。测试ema实现是否成功。
#
# ================================================================
import datetime
import json
from collections import deque
import paddle.fluid as fluid
import paddle.fluid.layers as P
import sys
import time
import shutil
import math
import copy
import random
import threading
import numpy as np
import os
from paddle.fluid.param_attr import ParamAttr
from paddle.fluid.param_attr import ParamAttr
from paddle.fluid.regularizer import L2Decay
from paddle.fluid.initializer import Constant
from paddle.fluid.regularizer import L2Decay
from paddle.fluid.layers.learning_rate_scheduler import _decay_step_counter
from paddle.fluid.optimizer import ExponentialMovingAverage
import paddle
import torch
from model.custom_layers import Conv2dUnit, DCNv2
from collections import OrderedDict
class MyNet(torch.nn.Module):
def __init__(self):
super(MyNet, self).__init__()
self.conv1 = torch.nn.Conv2d(3, 8, kernel_size=1, stride=1, padding=0, bias=True)
self.act1 = torch.nn.LeakyReLU(0.1)
self.dcnv2 = DCNv2(8, 512, filter_size=3, stride=2, padding=1, bias_attr=False)
def __call__(self, input_tensor):
x = self.conv1(input_tensor)
x = self.act1(x)
x = self.dcnv2(x)
return x
if __name__ == '__main__':
paddle.enable_static()
use_gpu = False
lr = 0.1
startup_prog = fluid.Program()
train_prog = fluid.Program()
with fluid.program_guard(train_prog, startup_prog):
with fluid.unique_name.guard():
inputs = P.data(name='input_1', shape=[-1, 3, 32, 32], append_batch_size=False, dtype='float32')
conv01_out_tensor = fluid.layers.conv2d(input=inputs, num_filters=8, filter_size=1, stride=1, padding=0,
param_attr=ParamAttr(name="conv01_weights"),
bias_attr=ParamAttr(name="conv01_bias", initializer=fluid.initializer.NormalInitializer(loc=0.0, scale=1.0, seed=0)))
act01_out_tensor = fluid.layers.leaky_relu(conv01_out_tensor, alpha=0.1)
filter_size = 3
filters = 512
stride = 2
padding = 1
conv_name = 'dcnv2'
offset_mask = fluid.layers.conv2d(
input=act01_out_tensor,
num_filters=filter_size * filter_size * 3,
filter_size=filter_size,
stride=stride,
padding=padding,
act=None,
# param_attr=ParamAttr(initializer=Constant(0.0), name=conv_name + "_conv_offset.w_0"),
param_attr=ParamAttr(name=conv_name + "_conv_offset.w_0"),
bias_attr=ParamAttr(initializer=Constant(0.0), name=conv_name + "_conv_offset.b_0"),
name=conv_name + "_conv_offset")
offset = offset_mask[:, :filter_size**2 * 2, :, :]
mask = offset_mask[:, filter_size**2 * 2:, :, :]
mask = fluid.layers.sigmoid(mask)
conv02_out_tensor = fluid.layers.deformable_conv(input=act01_out_tensor, offset=offset, mask=mask,
num_filters=filters,
filter_size=filter_size,
stride=stride,
padding=padding,
groups=1,
deformable_groups=1,
im2col_step=1,
param_attr=ParamAttr(name=conv_name + "_weights"),
bias_attr=False,
name=conv_name + ".conv2d.output.1")
# 建立损失函数
y_true = P.data(name='y_true', shape=[-1, 1, 16, 16], append_batch_size=False, dtype='float32')
# 先把差值逐项平方，可以用P.pow()这个op，也可以用python里的运算符**。
mseloss = P.pow(y_true - conv02_out_tensor, 2)
mseloss = P.reduce_mean(mseloss) # 再求平均，即mse损失函数
# 优化器
optimizer = fluid.optimizer.SGD(learning_rate=lr)
optimizer.minimize(mseloss)
eval_prog = fluid.Program()
with fluid.program_guard(eval_prog, startup_prog):
with fluid.unique_name.guard():
# 重新建立一次网络，用相同的张量名，不用写损失层
inputs = P.data(name='input_1', shape=[-1, 3, 32, 32], append_batch_size=False, dtype='float32')
conv01_out_tensor = fluid.layers.conv2d(input=inputs, num_filters=8, filter_size=1, stride=1, padding=0,
param_attr=ParamAttr(name="conv01_weights"),
bias_attr=ParamAttr(name="conv01_bias", initializer=fluid.initializer.NormalInitializer(loc=0.0, scale=1.0, seed=0)))
act01_out_tensor = fluid.layers.leaky_relu(conv01_out_tensor, alpha=0.1)
filter_size = 3
filters = 512
stride = 2
padding = 1
conv_name = 'dcnv2'
offset_mask = fluid.layers.conv2d(
input=act01_out_tensor,
num_filters=filter_size * filter_size * 3,
filter_size=filter_size,
stride=stride,
padding=padding,
act=None,
# param_attr=ParamAttr(initializer=Constant(0.0), name=conv_name + "_conv_offset.w_0"),
param_attr=ParamAttr(name=conv_name + "_conv_offset.w_0"),
bias_attr=ParamAttr(initializer=Constant(0.0), name=conv_name + "_conv_offset.b_0"),
name=conv_name + "_conv_offset")
offset = offset_mask[:, :filter_size**2 * 2, :, :]
mask = offset_mask[:, filter_size**2 * 2:, :, :]
mask = fluid.layers.sigmoid(mask)
conv02_out_tensor = fluid.layers.deformable_conv(input=act01_out_tensor, offset=offset, mask=mask,
num_filters=filters,
filter_size=filter_size,
stride=stride,
padding=padding,
groups=1,
deformable_groups=1,
im2col_step=1,
param_attr=ParamAttr(name=conv_name + "_weights"),
bias_attr=False,
name=conv_name + ".conv2d.output.1")
eval_fetch_list = [conv02_out_tensor]
eval_prog = eval_prog.clone(for_test=True)
# 参数初始化
gpu_id = int(os.environ.get('FLAGS_selected_gpus', 0))
place = fluid.CUDAPlace(gpu_id) if use_gpu else fluid.CPUPlace()
exe = fluid.Executor(place)
exe.run(startup_prog)
compiled_eval_prog = fluid.compiler.CompiledProgram(eval_prog)
# pytorch搭建的神经网络的权重。初始值是paddle相同层的初始值。为了模拟paddle训练过程。
# 1.卷积层
paddle_conv01_weights = np.array(fluid.global_scope().find_var('conv01_weights').get_tensor())
paddle_conv01_bias = np.array(fluid.global_scope().find_var('conv01_bias').get_tensor())
# 2.激活层
# 3.卷积层
paddle_conv02_weights = np.array(fluid.global_scope().find_var('dcnv2_conv_offset.w_0').get_tensor())
paddle_conv02_bias = np.array(fluid.global_scope().find_var('dcnv2_conv_offset.b_0').get_tensor())
paddle_conv02_dcn_weights = np.array(fluid.global_scope().find_var('dcnv2_weights').get_tensor())
# 6.激活层
# 7.损失函数层，没有权重。
# pytorch搭建的神经网络
myNet = MyNet()
# myNet = myNet.cuda()
loss_fn = torch.nn.MSELoss(reduce=True, size_average=True)
# loss_fn = loss_fn.cuda()
optimizer2 = torch.optim.SGD(filter(lambda p: p.requires_grad, myNet.parameters()), lr=lr) # requires_grad==True 的参数才可以被更新
# 初始化自己网络的权重
myNet.conv1.weight.data = torch.Tensor(np.copy(paddle_conv01_weights))
myNet.conv1.bias.data = torch.Tensor(np.copy(paddle_conv01_bias))
myNet.dcnv2.conv_offset.weight.data = torch.Tensor(np.copy(paddle_conv02_weights))
myNet.dcnv2.conv_offset.bias.data = torch.Tensor(np.copy(paddle_conv02_bias))
myNet.dcnv2.dcn_weight.data = torch.Tensor(np.copy(paddle_conv02_dcn_weights))
myNet.train() # 切换到训练模式
# 只训练8步
for step in range(8):
print('------------------ step %d ------------------' % step)
# ==================== train ====================
batch_data = np.random.normal(loc=0, scale=1, size=(2, 3, 32, 32)).astype(np.float32)
y_true_arr = np.random.normal(loc=0, scale=1, size=(2, 512, 16, 16)).astype(np.float32)
paddle_mseloss_out, paddle_conv01_out, paddle_conv02_out = exe.run(train_prog, feed={"input_1": batch_data, "y_true": y_true_arr, },
fetch_list=[mseloss, conv01_out_tensor, conv02_out_tensor])
print('train_forward:')
# python代码模拟训练过程，与paddle的输出校验。我们希望和飞桨有相同的输出。
batch_data = torch.Tensor(batch_data)
y_true_arr = torch.Tensor(y_true_arr)
my_act02_out = myNet(batch_data)
my_mseloss_out = loss_fn(my_act02_out, y_true_arr)
# 更新权重
optimizer2.zero_grad() # 清空上一步的残余更新参数值
my_mseloss_out.backward() # 误差反向传播, 计算参数更新值
optimizer2.step() # 将参数更新值施加到 net 的 parameters 上
_my_mseloss_out = my_mseloss_out.cpu().data.numpy()
diff_mseloss_out = np.sum((paddle_mseloss_out - _my_mseloss_out)**2)
print('diff_mseloss_out=%.6f' % diff_mseloss_out) # 若是0，则表示成功模拟出PaddlePaddle bn层的输出结果
print('\nbackward:')
# 和飞桨更新后的权重校验。
paddle_conv01_weights = np.array(fluid.global_scope().find_var('conv01_weights').get_tensor())
paddle_conv01_bias = np.array(fluid.global_scope().find_var('conv01_bias').get_tensor())
paddle_conv02_weights = np.array(fluid.global_scope().find_var('dcnv2_conv_offset.w_0').get_tensor())
paddle_conv02_bias = np.array(fluid.global_scope().find_var('dcnv2_conv_offset.b_0').get_tensor())
paddle_conv02_dcn_weights = np.array(fluid.global_scope().find_var('dcnv2_weights').get_tensor())
diff_conv01_weights = np.sum((paddle_conv01_weights - myNet.conv1.weight.data.numpy())**2)
print('diff_conv01_weights=%.6f' % diff_conv01_weights) # 若是0，则表示成功模拟出权重更新
diff_conv01_bias = np.sum((paddle_conv01_bias - myNet.conv1.bias.data.numpy())**2)
print('diff_conv01_bias=%.6f' % diff_conv01_bias) # 若是0，则表示成功模拟出权重更新
print('\nDCNv2:')
diff_conv02_weights = np.sum((paddle_conv02_weights - myNet.dcnv2.conv_offset.weight.data.numpy())**2)
print('diff_conv02_weights=%.6f' % diff_conv02_weights) # 若是0，则表示成功模拟出权重更新
diff_conv02_bias = np.sum((paddle_conv02_bias - myNet.dcnv2.conv_offset.bias.data.numpy())**2)
print('diff_conv02_bias=%.6f' % diff_conv02_bias) # 若是0，则表示成功模拟出权重更新
diff_conv02_dcn_weights = np.sum((paddle_conv02_dcn_weights - myNet.dcnv2.dcn_weight.data.numpy())**2)
print('diff_conv02_dcn_weights=%.6f' % diff_conv02_dcn_weights) # 若是0，则表示成功模拟出权重更新
# ==================== test ====================
test_data = np.random.normal(loc=0, scale=1, size=(2, 3, 32, 32)).astype(np.float32)
_conv02_out_tensor, = exe.run(compiled_eval_prog, feed={"input_1": test_data, }, fetch_list=eval_fetch_list)
# 自己网络的test
print('\ntest_forward:')
myNet.eval() # 切换到验证模式
test_data = torch.Tensor(test_data)
my_test_conv02_out_ = myNet(test_data)
my_test_conv02_out = my_test_conv02_out_.cpu().data.numpy()
myNet.train() # 切换到训练模式
d1 = np.sum((_conv02_out_tensor - my_test_conv02_out)**2)
print('d1=%.6f' % d1) # 若是0，则表示成功模拟出推理