Last active
July 21, 2020 05:43
-
-
Save phuocphn/6ef18c2c3a895d4545951c447f370f28 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import torch | |
import torch.nn as nn | |
from lsq import Conv2dLSQ, LinearLSQ, ActLSQ | |
__all__ = ['ResNet', 'resnet18', 'resnet34', 'resnet50', 'resnet101', | |
'resnet152', 'resnext50_32x4d', 'resnext101_32x8d', | |
'wide_resnet50_2', 'wide_resnet101_2'] | |
def conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1): | |
"""3x3 convolution with padding""" | |
return Conv2dLSQ(in_planes, out_planes, kernel_size=3, stride=stride, | |
padding=dilation, groups=groups, bias=False, dilation=dilation, nbits=8) | |
def conv1x1(in_planes, out_planes, stride=1): | |
"""1x1 convolution""" | |
return Conv2dLSQ(in_planes, out_planes, kernel_size=1, stride=stride, bias=False, nbits=8) | |
class BasicBlock(nn.Module): | |
expansion = 1 | |
def __init__(self, inplanes, planes, stride=1, downsample=None, groups=1, | |
base_width=64, dilation=1, norm_layer=None): | |
super(BasicBlock, self).__init__() | |
if norm_layer is None: | |
norm_layer = nn.BatchNorm2d | |
if groups != 1 or base_width != 64: | |
raise ValueError('BasicBlock only supports groups=1 and base_width=64') | |
if dilation > 1: | |
raise NotImplementedError("Dilation > 1 not supported in BasicBlock") | |
# Both self.conv1 and self.downsample layers downsample the input when stride != 1 | |
self.conv1 = conv3x3(inplanes, planes, stride) | |
self.bn1 = norm_layer(planes) | |
self.relu = nn.ReLU(inplace=True) | |
self.act1 = ActLSQ(nbits=8) | |
self.conv2 = conv3x3(planes, planes) | |
self.bn2 = norm_layer(planes) | |
self.act2 = ActLSQ(nbits=8) | |
self.downsample = downsample | |
self.stride = stride | |
def forward(self, x): | |
identity = x | |
out = self.conv1(x) | |
out = self.bn1(out) | |
out = self.relu(out) | |
out = self.act1(out) | |
out = self.conv2(out) | |
out = self.bn2(out) | |
if self.downsample is not None: | |
identity = self.downsample(x) | |
out += identity | |
out = self.relu(out) | |
out = self.act2(out) | |
return out | |
class ResNet(nn.Module): | |
def __init__(self, block, layers, num_classes=1000, zero_init_residual=False, | |
groups=1, width_per_group=64, replace_stride_with_dilation=None, | |
norm_layer=None): | |
super(ResNet, self).__init__() | |
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 " | |
"or a 3-element tuple, got {}".format(replace_stride_with_dilation)) | |
self.groups = groups | |
self.base_width = width_per_group | |
self.act0 = ActLSQ(nbits=8, signed=True) | |
self.conv1 = Conv2dLSQ(3, self.inplanes, kernel_size=7, stride=2, padding=3, | |
bias=False, nbits=8) | |
self.bn1 = norm_layer(self.inplanes) | |
self.relu = nn.ReLU(inplace=True) | |
self.act1 = ActLSQ(nbits=8) | |
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.act2 = ActLSQ(nbits=8) | |
self.avgpool = nn.AdaptiveAvgPool2d((1, 1)) | |
self.fc = LinearLSQ(512 * block.expansion, num_classes, nbits=8) | |
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): | |
nn.init.constant_(m.bn3.weight, 0) | |
elif isinstance(m, BasicBlock): | |
nn.init.constant_(m.bn2.weight, 0) | |
def _make_layer(self, block, planes, blocks, stride=1, dilate=False): | |
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): | |
# See note [TorchScript super()] | |
x= self.act0(x) | |
x = self.conv1(x) | |
x = self.bn1(x) | |
x = self.relu(x) | |
x = self.act1(x) | |
x = self.maxpool(x) | |
x = self.layer1(x) | |
x = self.layer2(x) | |
x = self.layer3(x) | |
x = self.layer4(x) | |
x = self.act2(x) | |
x = self.avgpool(x) | |
x = torch.flatten(x, 1) | |
x = self.fc(x) | |
return x | |
def forward(self, x): | |
return self._forward_impl(x) | |
model_urls = { | |
'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth', | |
'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth', | |
'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth', | |
'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth', | |
'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth', | |
'resnext50_32x4d': 'https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pth', | |
'resnext101_32x8d': 'https://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pth', | |
'wide_resnet50_2': 'https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pth', | |
'wide_resnet101_2': 'https://download.pytorch.org/models/wide_resnet101_2-32ee1156.pth', | |
} | |
def _resnet(arch, block, layers, pretrained, progress, **kwargs): | |
model = ResNet(block, layers, **kwargs) | |
from torch.utils.model_zoo import load_url as load_state_dict_from_url | |
if pretrained: | |
model_dict = model.state_dict() | |
state_dict = load_state_dict_from_url(model_urls[arch], | |
progress=progress) | |
ignore_keys = ["act0.alpha", "act0.init_state", "conv1.alpha", | |
"conv1.init_state", "act1.alpha", "act1.init_state", | |
"layer1.0.conv1.alpha", "layer1.0.conv1.init_state", | |
"layer1.0.act1.alpha", "layer1.0.act1.init_state", | |
"layer1.0.conv2.alpha", "layer1.0.conv2.init_state", | |
"layer1.0.act2.alpha", "layer1.0.act2.init_state", | |
"layer1.1.conv1.alpha", "layer1.1.conv1.init_state", | |
"layer1.1.act1.alpha", "layer1.1.act1.init_state", | |
"layer1.1.conv2.alpha", "layer1.1.conv2.init_state", | |
"layer1.1.act2.alpha", "layer1.1.act2.init_state", | |
"layer2.0.conv1.alpha", "layer2.0.conv1.init_state", | |
"layer2.0.act1.alpha", "layer2.0.act1.init_state", | |
"layer2.0.conv2.alpha", "layer2.0.conv2.init_state", | |
"layer2.0.act2.alpha", "layer2.0.act2.init_state", "layer2.0.downsample.0.alpha", | |
"layer2.0.downsample.0.init_state", "layer2.1.conv1.alpha", "layer2.1.conv1.init_state", | |
"layer2.1.act1.alpha", "layer2.1.act1.init_state", "layer2.1.conv2.alpha", | |
"layer2.1.conv2.init_state", "layer2.1.act2.alpha", "layer2.1.act2.init_state", | |
"layer3.0.conv1.alpha", "layer3.0.conv1.init_state", "layer3.0.act1.alpha", | |
"layer3.0.act1.init_state", "layer3.0.conv2.alpha", "layer3.0.conv2.init_state", | |
"layer3.0.act2.alpha", "layer3.0.act2.init_state", "layer3.0.downsample.0.alpha", | |
"layer3.0.downsample.0.init_state", "layer3.1.conv1.alpha", "layer3.1.conv1.init_state", | |
"layer3.1.act1.alpha", "layer3.1.act1.init_state", "layer3.1.conv2.alpha", | |
"layer3.1.conv2.init_state", "layer3.1.act2.alpha", "layer3.1.act2.init_state", | |
"layer4.0.conv1.alpha", "layer4.0.conv1.init_state", "layer4.0.act1.alpha", | |
"layer4.0.act1.init_state", "layer4.0.conv2.alpha", "layer4.0.conv2.init_state", | |
"layer4.0.act2.alpha", "layer4.0.act2.init_state", "layer4.0.downsample.0.alpha", | |
"layer4.0.downsample.0.init_state", "layer4.1.conv1.alpha", "layer4.1.conv1.init_state", | |
"layer4.1.act1.alpha", "layer4.1.act1.init_state", "layer4.1.conv2.alpha", | |
"layer4.1.conv2.init_state", "layer4.1.act2.alpha", "layer4.1.act2.init_state", | |
"fc.init_state", "fc.alpha"] | |
s = { k:v for k, v in state_dict.items() if k not in ignore_keys} | |
model_dict.update(s) | |
model.load_state_dict(model_dict) | |
return model | |
def resnet18(pretrained=False, progress=True, **kwargs): | |
r"""ResNet-18 model from | |
`"Deep Residual Learning for Image Recognition" <https://arxiv.org/pdf/1512.03385.pdf>`_ | |
Args: | |
pretrained (bool): If True, returns a model pre-trained on ImageNet | |
progress (bool): If True, displays a progress bar of the download to stderr | |
""" | |
return _resnet('resnet18', BasicBlock, [2, 2, 2, 2], pretrained, progress, | |
**kwargs) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment