Spaces:
Runtime error
Runtime error
''' | |
For MEMO implementations of ImageNet-ResNet | |
Reference: | |
https://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py | |
''' | |
import torch | |
import torch.nn as nn | |
try: | |
from torchvision.models.utils import load_state_dict_from_url | |
except: | |
from torch.hub import load_state_dict_from_url | |
__all__ = ['ResNet', 'resnet18', 'resnet34', 'resnet50', 'resnet101', | |
'resnet152', 'resnext50_32x4d', 'resnext101_32x8d', | |
'wide_resnet50_2', 'wide_resnet101_2'] | |
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 conv3x3(in_planes, out_planes, stride=1, groups=1, dilation=1): | |
"""3x3 convolution with padding""" | |
return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, | |
padding=dilation, groups=groups, bias=False, dilation=dilation) | |
def conv1x1(in_planes, out_planes, stride=1): | |
"""1x1 convolution""" | |
return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False) | |
class BasicBlock(nn.Module): | |
expansion = 1 | |
__constants__ = ['downsample'] | |
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.conv2 = conv3x3(planes, planes) | |
self.bn2 = norm_layer(planes) | |
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.conv2(out) | |
out = self.bn2(out) | |
if self.downsample is not None: | |
identity = self.downsample(x) | |
out += identity | |
out = self.relu(out) | |
return out | |
class Bottleneck(nn.Module): | |
expansion = 4 | |
__constants__ = ['downsample'] | |
def __init__(self, inplanes, planes, stride=1, downsample=None, groups=1, | |
base_width=64, dilation=1, norm_layer=None): | |
super(Bottleneck, self).__init__() | |
if norm_layer is None: | |
norm_layer = nn.BatchNorm2d | |
width = int(planes * (base_width / 64.)) * groups | |
# Both self.conv2 and self.downsample layers downsample the input when stride != 1 | |
self.conv1 = conv1x1(inplanes, width) | |
self.bn1 = norm_layer(width) | |
self.conv2 = conv3x3(width, width, stride, groups, dilation) | |
self.bn2 = norm_layer(width) | |
self.conv3 = conv1x1(width, planes * self.expansion) | |
self.bn3 = norm_layer(planes * self.expansion) | |
self.relu = nn.ReLU(inplace=True) | |
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.conv2(out) | |
out = self.bn2(out) | |
out = self.relu(out) | |
out = self.conv3(out) | |
out = self.bn3(out) | |
if self.downsample is not None: | |
identity = self.downsample(x) | |
out += identity | |
out = self.relu(out) | |
return out | |
class GeneralizedResNet_imagenet(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(GeneralizedResNet_imagenet, 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.conv1 = nn.Conv2d(3, self.inplanes, kernel_size=7, stride=2, padding=3, # stride=2 -> stride=1 for cifar | |
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) # Removed in _forward_impl for cifar | |
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.out_dim = 512 * block.expansion | |
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): | |
x = self.conv1(x) | |
x = self.bn1(x) | |
x = self.relu(x) | |
x = self.maxpool(x) | |
x_1 = self.layer1(x) | |
x_2 = self.layer2(x_1) | |
x_3 = self.layer3(x_2) | |
return x_3 | |
def forward(self, x): | |
return self._forward_impl(x) | |
class SpecializedResNet_imagenet(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(SpecializedResNet_imagenet, self).__init__() | |
if norm_layer is None: | |
norm_layer = nn.BatchNorm2d | |
self._norm_layer = norm_layer | |
self.feature_dim = 512 * block.expansion | |
self.inplanes = 256 * block.expansion | |
self.dilation = 1 | |
if replace_stride_with_dilation is None: | |
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.layer4 = self._make_layer(block, 512, layers[3], stride=2, | |
dilate=replace_stride_with_dilation[2]) | |
self.avgpool = nn.AdaptiveAvgPool2d((1, 1)) | |
self.out_dim = 512 * block.expansion | |
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) | |
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(self,x): | |
x_4 = self.layer4(x) # [bs, 512, 4, 4] | |
pooled = self.avgpool(x_4) # [bs, 512, 1, 1] | |
features = torch.flatten(pooled, 1) # [bs, 512] | |
return features | |
def get_resnet10_imagenet(): | |
basenet = GeneralizedResNet_imagenet(BasicBlock,[1, 1, 1, 1]) | |
adaptivenet = SpecializedResNet_imagenet(BasicBlock, [1, 1, 1, 1]) | |
return basenet,adaptivenet | |
def get_resnet18_imagenet(): | |
basenet = GeneralizedResNet_imagenet(BasicBlock,[2, 2, 2, 2]) | |
adaptivenet = SpecializedResNet_imagenet(BasicBlock, [2, 2, 2, 2]) | |
return basenet,adaptivenet | |
def get_resnet26_imagenet(): | |
basenet = GeneralizedResNet_imagenet(Bottleneck,[2, 2, 2, 2]) | |
adaptivenet = SpecializedResNet_imagenet(Bottleneck, [2, 2, 2, 2]) | |
return basenet,adaptivenet | |
def get_resnet34_imagenet(): | |
basenet = GeneralizedResNet_imagenet(BasicBlock,[3, 4, 6, 3]) | |
adaptivenet = SpecializedResNet_imagenet(BasicBlock, [3, 4, 6, 3]) | |
return basenet,adaptivenet | |
def get_resnet50_imagenet(): | |
basenet = GeneralizedResNet_imagenet(Bottleneck,[3, 4, 6, 3]) | |
adaptivenet = SpecializedResNet_imagenet(Bottleneck, [3, 4, 6, 3]) | |
return basenet,adaptivenet | |
if __name__ == '__main__': | |
model2imagenet = 3*224*224 | |
a, b = get_resnet10_imagenet() | |
_base = sum(p.numel() for p in a.parameters()) | |
_adap = sum(p.numel() for p in b.parameters()) | |
print(f"resnet10 #params:{_base+_adap}") | |
a, b = get_resnet18_imagenet() | |
_base = sum(p.numel() for p in a.parameters()) | |
_adap = sum(p.numel() for p in b.parameters()) | |
print(f"resnet18 #params:{_base+_adap}") | |
a, b = get_resnet26_imagenet() | |
_base = sum(p.numel() for p in a.parameters()) | |
_adap = sum(p.numel() for p in b.parameters()) | |
print(f"resnet26 #params:{_base+_adap}") | |
a, b = get_resnet34_imagenet() | |
_base = sum(p.numel() for p in a.parameters()) | |
_adap = sum(p.numel() for p in b.parameters()) | |
print(f"resnet34 #params:{_base+_adap}") | |
a, b = get_resnet50_imagenet() | |
_base = sum(p.numel() for p in a.parameters()) | |
_adap = sum(p.numel() for p in b.parameters()) | |
print(f"resnet50 #params:{_base+_adap}") |