import math import torch.nn as nn from .lib.nn import SynchronizedBatchNorm2d from .utils import load_url BatchNorm2d = SynchronizedBatchNorm2d __all__ = ["ResNeXt", "resnext101"] # support resnext 101 model_urls = { #'resnext50': 'http://sceneparsing.csail.mit.edu/model/pretrained_resnet/resnext50-imagenet.pth', "resnext101": "http://sceneparsing.csail.mit.edu/model/pretrained_resnet/resnext101-imagenet.pth" } def conv3x3(in_planes, out_planes, stride=1): "3x3 convolution with padding" return nn.Conv2d( in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False ) class GroupBottleneck(nn.Module): expansion = 2 def __init__(self, inplanes, planes, stride=1, groups=1, downsample=None): super(GroupBottleneck, self).__init__() self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False) self.bn1 = BatchNorm2d(planes) self.conv2 = nn.Conv2d( planes, planes, kernel_size=3, stride=stride, padding=1, groups=groups, bias=False, ) self.bn2 = BatchNorm2d(planes) self.conv3 = nn.Conv2d(planes, planes * 2, kernel_size=1, bias=False) self.bn3 = BatchNorm2d(planes * 2) self.relu = nn.ReLU(inplace=True) self.downsample = downsample self.stride = stride def forward(self, x): residual = 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: residual = self.downsample(x) out += residual out = self.relu(out) return out class ResNeXt(nn.Module): def __init__(self, block, layers, groups=32, num_classes=1000): self.inplanes = 128 super(ResNeXt, self).__init__() self.conv1 = conv3x3(3, 64, stride=2) self.bn1 = BatchNorm2d(64) self.relu1 = nn.ReLU(inplace=True) self.conv2 = conv3x3(64, 64) self.bn2 = BatchNorm2d(64) self.relu2 = nn.ReLU(inplace=True) self.conv3 = conv3x3(64, 128) self.bn3 = BatchNorm2d(128) self.relu3 = nn.ReLU(inplace=True) self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1) self.layer1 = self._make_layer(block, 128, layers[0], groups=groups) self.layer2 = self._make_layer(block, 256, layers[1], stride=2, groups=groups) self.layer3 = self._make_layer(block, 512, layers[2], stride=2, groups=groups) self.layer4 = self._make_layer(block, 1024, layers[3], stride=2, groups=groups) self.avgpool = nn.AvgPool2d(7, stride=1) self.fc = nn.Linear(1024 * block.expansion, num_classes) for m in self.modules(): if isinstance(m, nn.Conv2d): n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels // m.groups m.weight.data.normal_(0, math.sqrt(2.0 / n)) elif isinstance(m, BatchNorm2d): m.weight.data.fill_(1) m.bias.data.zero_() def _make_layer(self, block, planes, blocks, stride=1, groups=1): downsample = None if stride != 1 or self.inplanes != planes * block.expansion: downsample = nn.Sequential( nn.Conv2d( self.inplanes, planes * block.expansion, kernel_size=1, stride=stride, bias=False, ), BatchNorm2d(planes * block.expansion), ) layers = [] layers.append(block(self.inplanes, planes, stride, groups, downsample)) self.inplanes = planes * block.expansion for i in range(1, blocks): layers.append(block(self.inplanes, planes, groups=groups)) return nn.Sequential(*layers) def forward(self, x): x = self.relu1(self.bn1(self.conv1(x))) x = self.relu2(self.bn2(self.conv2(x))) x = self.relu3(self.bn3(self.conv3(x))) x = self.maxpool(x) x = self.layer1(x) x = self.layer2(x) x = self.layer3(x) x = self.layer4(x) x = self.avgpool(x) x = x.view(x.size(0), -1) x = self.fc(x) return x ''' def resnext50(pretrained=False, **kwargs): """Constructs a ResNet-50 model. Args: pretrained (bool): If True, returns a model pre-trained on Places """ model = ResNeXt(GroupBottleneck, [3, 4, 6, 3], **kwargs) if pretrained: model.load_state_dict(load_url(model_urls['resnext50']), strict=False) return model ''' def resnext101(pretrained=False, **kwargs): """Constructs a ResNet-101 model. Args: pretrained (bool): If True, returns a model pre-trained on Places """ model = ResNeXt(GroupBottleneck, [3, 4, 23, 3], **kwargs) if pretrained: model.load_state_dict(load_url(model_urls["resnext101"]), strict=False) return model # def resnext152(pretrained=False, **kwargs): # """Constructs a ResNeXt-152 model. # # Args: # pretrained (bool): If True, returns a model pre-trained on Places # """ # model = ResNeXt(GroupBottleneck, [3, 8, 36, 3], **kwargs) # if pretrained: # model.load_state_dict(load_url(model_urls['resnext152'])) # return model