-
Notifications
You must be signed in to change notification settings - Fork 0
/
resnet18.py
354 lines (304 loc) · 11.8 KB
/
resnet18.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
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
import torch
import torchvision
from torchvision import datasets
from torch import nn
from torch.utils.data import DataLoader
import time
from torch import Tensor
from torch.utils.tensorboard import SummaryWriter
from torchvision.models.resnet import BasicBlock, Bottleneck, conv1x1, conv3x3
from functools import partial
from typing import Type, Any, Callable, Union, List, Optional
writer = SummaryWriter('runs/resnet50_cifar10')
torch.cuda.set_device(1)
# Download training data from open datasets.
training_data = datasets.CIFAR10(
root="data",
train=True,
download=True,
transform=torchvision.transforms.ToTensor(),
)
# Download test data from open datasets.
test_data = datasets.CIFAR10(
root="data",
train=False,
download=True,
transform=torchvision.transforms.ToTensor(),
)
batch_size = 400
dropout_p = 0.5
# Create data loaders.
train_dataloader = DataLoader(training_data, batch_size=batch_size)
test_dataloader = DataLoader(test_data, batch_size=batch_size)
input_shape = None
for X, y in test_dataloader:
input_shape = X.shape
print(f"Shape of X [N, C, H, W]: {X.shape}")
print(f"Shape of y: {y.shape} {y.dtype}")
break
device = "cuda"
def get_lr(optimizer):
for param_group in optimizer.param_groups:
return param_group['lr']
@torch.no_grad()
def init_weight(m):
if isinstance(m, torch.nn.Conv2d) or isinstance(m, torch.nn.Linear):
torch.nn.init.xavier_uniform_(m.weight)
if isinstance(m, torch.nn.Linear):
m.bias.data.fill_(0)
class MyResNet(nn.Module):
def __init__(
self,
block: Type[Union[BasicBlock, Bottleneck]],
layers: List[int],
num_classes: int = 1000,
zero_init_residual: bool = False,
groups: int = 1,
width_per_group: int = 64,
replace_stride_with_dilation: Optional[List[bool]] = None,
norm_layer: Optional[Callable[..., nn.Module]] = None,
) -> None:
super().__init__()
# _log_api_usage_once(self)
if norm_layer is None:
norm_layer = nn.BatchNorm2d
self._norm_layer = norm_layer
self.inplanes = 64
self.dilation = 1
if replace_stride_with_dilation is None:
# each element in the tuple indicates if we should replace
# the 2x2 stride with a dilated convolution instead
replace_stride_with_dilation = [False, False, False]
if len(replace_stride_with_dilation) != 3:
raise ValueError(
"replace_stride_with_dilation should be None "
f"or a 3-element tuple, got {replace_stride_with_dilation}"
)
self.groups = groups
self.base_width = width_per_group
self.conv1 = nn.Conv2d(
3, self.inplanes, kernel_size=7, stride=2, padding=3, bias=False)
self.bn1 = norm_layer(self.inplanes)
self.relu = nn.ReLU(inplace=True)
self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(
block, 128, layers[1], stride=2, dilate=replace_stride_with_dilation[0])
self.layer3 = self._make_layer(
block, 256, layers[2], stride=2, dilate=replace_stride_with_dilation[1])
self.layer4 = self._make_layer(
block, 512, layers[3], stride=2, dilate=replace_stride_with_dilation[2])
self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
self.fc = nn.Linear(512 * block.expansion, num_classes)
self.drop1 = nn.Dropout(dropout_p)
self.drop2 = nn.Dropout(dropout_p)
self.drop3 = nn.Dropout(dropout_p)
self.drop4 = nn.Dropout(dropout_p)
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(
m.weight, mode="fan_out", nonlinearity="relu")
elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
# Zero-initialize the last BN in each residual branch,
# so that the residual branch starts with zeros, and each residual block behaves like an identity.
# This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677
if zero_init_residual:
for m in self.modules():
if isinstance(m, Bottleneck) and m.bn3.weight is not None:
# type: ignore[arg-type]
nn.init.constant_(m.bn3.weight, 0)
elif isinstance(m, BasicBlock) and m.bn2.weight is not None:
# type: ignore[arg-type]
nn.init.constant_(m.bn2.weight, 0)
def _make_layer(
self,
block: Type[Union[BasicBlock, Bottleneck]],
planes: int,
blocks: int,
stride: int = 1,
dilate: bool = False,
) -> nn.Sequential:
norm_layer = self._norm_layer
downsample = None
previous_dilation = self.dilation
if dilate:
self.dilation *= stride
stride = 1
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = nn.Sequential(
conv1x1(self.inplanes, planes * block.expansion, stride),
norm_layer(planes * block.expansion),
)
layers = []
layers.append(
block(
self.inplanes, planes, stride, downsample, self.groups, self.base_width, previous_dilation, norm_layer
)
)
self.inplanes = planes * block.expansion
for _ in range(1, blocks):
layers.append(
block(
self.inplanes,
planes,
groups=self.groups,
base_width=self.base_width,
dilation=self.dilation,
norm_layer=norm_layer,
)
)
return nn.Sequential(*layers)
def _forward_impl(self, x: Tensor) -> Tensor:
# See note [TorchScript super()]
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.maxpool(x)
x = self.layer1(x)
x = self.drop1(x)
x = self.layer2(x)
x = self.drop2(x)
x = self.layer3(x)
x = self.drop3(x)
x = self.layer4(x)
x = self.drop4(x)
x = self.avgpool(x)
x = torch.flatten(x, 1)
x = self.fc(x)
return x
def forward(self, x: Tensor) -> Tensor:
return self._forward_impl(x)
class Cifar10Model(nn.Module):
def __init__(self, num_classes):
super().__init__()
# input (batch, 3, 32, 32)
self.conv1 = nn.Conv2d(3, 8, kernel_size=7, stride=2, padding=3, bias=False)
self.bn1 = nn.BatchNorm2d(8)
self.relu1 = nn.ReLU(inplace=True)
# (batch, 8, 16, 16)
self.conv2 = nn.Conv2d(8, 16, kernel_size=3, stride=2, padding=1)
self.bn2 = nn.BatchNorm2d(16)
self.relu2 = nn.ReLU(inplace=True)
# (batch, 16, 7, 7)
self.conv3 = nn.Conv2d(16, 32, kernel_size=3, stride=2, padding=1)
self.bn3 = nn.BatchNorm2d(32)
self.relu3 = nn.ReLU(inplace=True)
# (batch, 32, 3, 3)
self.conv4 = nn.Conv2d(32, 64, kernel_size=3, stride=2, padding=1)
self.bn4 = nn.BatchNorm2d(64)
self.relu4 = nn.ReLU(inplace=True)
# (batch, 64, 2, 2)
self.flat = nn.Flatten()
self.linear = nn.Linear(256, num_classes)
def forward(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu1(x)
x = self.conv2(x)
x = self.bn2(x)
x = self.relu2(x)
x = self.conv3(x)
x = self.bn3(x)
x = self.relu3(x)
x = self.conv4(x)
x = self.bn4(x)
x = self.relu4(x)
x = self.flat(x)
x = self.linear(x)
return x
# model = MyResNet(block=torchvision.models.resnet.BasicBlock, layers=[1,1,1,1], num_classes=10)
model = Cifar10Model(10)
model.apply(init_weight)
model.to(device)
epochs = 500
loss_fn = nn.CrossEntropyLoss()
lrs = []
clip_value = 0.1
max_lr = 0.01
weight_decay = 1e-4
optimizer = torch.optim.Adam(
model.parameters(), max_lr, weight_decay=weight_decay)
sched = torch.optim.lr_scheduler.OneCycleLR(optimizer, max_lr, epochs=epochs,
steps_per_epoch=len(train_dataloader))
def train(dataloader, model, loss_fn, optimizer, t):
size = len(dataloader.dataset)
model.train()
for batch, (X, y) in enumerate(dataloader):
start_time = time.time()
X, y = X.to(device), y.to(device)
# Compute prediction error
pred = model(X)
loss = loss_fn(pred, y)
# Backpropagation
loss.backward()
torch.nn.utils.clip_grad_value_(model.parameters(), clip_value)
optimizer.step()
optimizer.zero_grad()
lrs.append(get_lr(optimizer))
sched.step()
# if batch % 10 == 0:
# loss, current = loss.item(), batch * len(X)
# cost_time = time.time() - start_time
# print(f"Train: \n loss: {loss:>7f} [{current:>5d}/{size:>5d}] cost: {cost_time:>4f}")
# writer.add_scalar("train_loss", loss,
# t * len(dataloader) + batch)
def test(dataloader, model, loss_fn, t):
size = len(dataloader.dataset)
num_batches = len(dataloader)
model.eval()
test_loss, correct = 0, 0
with torch.no_grad():
for X, y in dataloader:
X, y = X.to(device), y.to(device)
pred = model(X)
test_loss += loss_fn(pred, y).item()
correct += (pred.argmax(1) == y).type(torch.float).sum().item()
test_loss /= num_batches
correct /= size
writer.add_scalar("test_acc", 100*correct, t)
print(
f"Test: \n Accuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f} \n")
def test2(dataloader, model, loss_fn, t):
size = len(dataloader.dataset)
num_batches = len(dataloader)
model.eval()
test_loss, correct = 0, 0
with torch.no_grad():
for X, y in dataloader:
X, y = X.to(device), y.to(device)
pred = model(X)
test_loss += loss_fn(pred, y).item()
correct += (pred.argmax(1) == y).type(torch.float).sum().item()
test_loss /= num_batches
correct /= size
writer.add_scalar("test_acc", 100*correct, t)
print(
f"Train Test: \n Accuracy: {(100*correct):>0.1f}%, Avg loss: {test_loss:>8f} \n")
for t in range(epochs):
print(f"Epoch {t+1}\n-------------------------------")
train(train_dataloader, model, loss_fn, optimizer, t)
test2(train_dataloader, model, loss_fn, t)
test(test_dataloader, model, loss_fn, t)
# print("lrs:")
# for iter in lrs:
# print(iter, sep=",")
# print("")
print("Done!")
model.to("cpu")
torch.save(model, "resnet9.pt")
# Export the model
dummy_input = torch.randn(input_shape)
torch.onnx.export(model, # model being run
# model input (or a tuple for multiple inputs)
dummy_input,
# where to save the model (can be a file or file-like object)
"resnet9.onnx",
export_params=True, # store the trained parameter weights inside the model file
# opset_version=10, # the ONNX version to export the model to
do_constant_folding=True, # whether to execute constant folding for optimization
input_names=['input'], # the model's input names
output_names=['output'], # the model's output names
dynamic_axes={'input': {0: 'batch_size'}, # variable length axes
'output': {0: 'batch_size'}})