Source code for backbone.ResNetBottleneck

import torch
from torch import Tensor
import torch.nn as nn
from typing import Type, Any, Callable, Union, List, Optional

from backbone import MammothBackbone, register_backbone


__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-f37072fd.pth',
    'resnet34': 'https://download.pytorch.org/models/resnet34-b627a593.pth',
    'resnet50': 'https://download.pytorch.org/models/resnet50-0676ba61.pth',
    'resnet101': 'https://download.pytorch.org/models/resnet101-63fe2227.pth',
    'resnet152': 'https://download.pytorch.org/models/resnet152-394f9c45.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',
}


[docs] def conv3x3(in_planes: int, out_planes: int, stride: int = 1, groups: int = 1, dilation: int = 1) -> nn.Conv2d: """3x3 convolution with padding""" return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride, padding=dilation, groups=groups, bias=False, dilation=dilation)
[docs] def conv1x1(in_planes: int, out_planes: int, stride: int = 1) -> nn.Conv2d: """1x1 convolution""" return nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)
[docs] class Bottleneck(nn.Module): # Bottleneck in torchvision places the stride for downsampling at 3x3 convolution(self.conv2) # while original implementation places the stride at the first 1x1 convolution(self.conv1) # according to "Deep residual learning for image recognition"https://arxiv.org/abs/1512.03385. # This variant is also known as ResNet V1.5 and improves accuracy according to # https://ngc.nvidia.com/catalog/model-scripts/nvidia:resnet_50_v1_5_for_pytorch. expansion: int = 4 def __init__( self, inplanes: int, planes: int, stride: int = 1, downsample: Optional[nn.Module] = None, groups: int = 1, base_width: int = 64, dilation: int = 1, norm_layer: Optional[Callable[..., nn.Module]] = None ) -> None: super(Bottleneck, self).__init__() self.return_prerelu = False 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
[docs] def forward(self, x: Tensor) -> Tensor: 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 if self.return_prerelu: self.prerelu = out.clone() out = self.relu(out) return out
[docs] class ResNet(MammothBackbone): def __init__( self, block: Bottleneck, layers: List[int], num_classes: int = 1000, zero_init_residual: bool = False, groups: int = 1, pretrained: bool = False, width_per_group: int = 64, replace_stride_with_dilation: Optional[List[bool]] = None, norm_layer: Optional[Callable[..., nn.Module]] = None ) -> None: super(ResNet, self).__init__() self.block = block self.device = "cpu" if norm_layer is None: norm_layer = nn.BatchNorm2d self._norm_layer = norm_layer self.return_prerelu = False self.num_classes = num_classes 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, 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.classifier = nn.Linear(512 * block.expansion, num_classes) self.feature_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): # type: ignore[arg-type] nn.init.constant_(m.bn3.weight, 0) if pretrained: ckpt = torch.hub.load_state_dict_from_url(model_urls['resnet50'], progress=True, check_hash=True) # drop classifier weights ckpt.pop('fc.weight') ckpt.pop('fc.bias') missing, unexpected = self.load_state_dict(ckpt, strict=False) assert len([x for x in missing if 'classifier' not in x]) == 0, "Missing keys: {}".format(missing) assert len(unexpected) == 0, "Unexpected keys: {}".format(unexpected)
[docs] def to(self, device, **kwargs): self.device = device return super().to(device, **kwargs)
[docs] def set_return_prerelu(self, enable=True): self.return_prerelu = enable for c in self.modules(): if isinstance(c, self.block): c.return_prerelu = enable
def _make_layer(self, block: 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)
[docs] def forward(self, x: Tensor, returnt="out") -> Tensor: out_0 = self.conv1(x) out_0 = self.bn1(out_0) if self.return_prerelu: out_0_t = out_0.clone() out_0 = self.relu(out_0) out_0 = self.maxpool(out_0) out_1 = self.layer1(out_0) out_2 = self.layer2(out_1) out_3 = self.layer3(out_2) out_4 = self.layer4(out_3) feature = self.avgpool(out_4) feature = torch.flatten(feature, 1) if returnt == 'features': return feature out = self.classifier(feature) if returnt == 'out': return out elif returnt == 'full': return out, [ out_0 if not self.return_prerelu else out_0_t, out_1 if not self.return_prerelu else self.layer1[-1].prerelu, out_2 if not self.return_prerelu else self.layer2[-1].prerelu, out_3 if not self.return_prerelu else self.layer3[-1].prerelu, out_4 if not self.return_prerelu else self.layer4[-1].prerelu ] elif returnt == 'both': return (out, feature) raise NotImplementedError("Unknown return type. Must be in ['out', 'features', 'both', 'full'] but got {}".format(returnt))
[docs] def set_grad_filter(self, filter_s: str, enable: bool) -> None: negative_mode = filter_s[0] == '~' if negative_mode: filter_s = filter_s[1:] for _, p in filter(lambda x: filter_s not in x[0], self.named_parameters()): p.requires_grad = enable else: for _, p in filter(lambda x: filter_s in x[0], self.named_parameters()): p.requires_grad = enable
@register_backbone("resnet50") def resnet50(num_classes: int, pretrained=False) -> ResNet: r"""ResNet-50 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(Bottleneck, [3, 4, 6, 3], num_classes=num_classes, pretrained=pretrained)
[docs] @register_backbone("resnet50_pt") def resnet50(num_classes: int) -> ResNet: r"""ResNet-50 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(Bottleneck, [3, 4, 6, 3], num_classes=num_classes, pretrained=True)