Shortcuts

Source code for tllib.vision.models.segmentation.deeplabv2

"""
@author: Junguang Jiang
@contact: JiangJunguang1123@outlook.com
"""
import torch.nn as nn
from torchvision.models.utils import load_state_dict_from_url


model_urls = {
    'deeplabv2_resnet101': 'https://cloud.tsinghua.edu.cn/f/2d9a7fc43ce34f76803a/?dl=1'
}

affine_par = True


class Bottleneck(nn.Module):
    expansion = 4

    def __init__(self, inplanes, planes, stride=1, dilation=1, downsample=None):
        super(Bottleneck, self).__init__()
        self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, stride=stride, bias=False)  # change
        self.bn1 = nn.BatchNorm2d(planes, affine=affine_par)
        for i in self.bn1.parameters():
            i.requires_grad = False

        padding = dilation

        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1,  # change
                               padding=padding, bias=False, dilation=dilation)
        self.bn2 = nn.BatchNorm2d(planes, affine=affine_par)
        for i in self.bn2.parameters():
            i.requires_grad = False

        self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
        self.bn3 = nn.BatchNorm2d(planes * 4, affine=affine_par)
        for i in self.bn3.parameters():
            i.requires_grad = False

        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 ASPP_V2(nn.Module):
    def __init__(self, inplanes, dilation_series, padding_series, num_classes):
        super(ASPP_V2, self).__init__()
        self.conv2d_list = nn.ModuleList()
        for dilation, padding in zip(dilation_series, padding_series):
            self.conv2d_list.append(
                nn.Conv2d(inplanes, num_classes, kernel_size=3, stride=1, padding=padding, dilation=dilation,
                          bias=True))
        for m in self.conv2d_list:
            m.weight.data.normal_(0, 0.01)

    def forward(self, x):
        out = self.conv2d_list[0](x)
        for i in range(len(self.conv2d_list) - 1):
            out += self.conv2d_list[i + 1](x)

        return out


class ResNet(nn.Module):
    def __init__(self, block, layers):
        self.inplanes = 64
        super(ResNet, self).__init__()
        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
                               bias=False)
        self.bn1 = nn.BatchNorm2d(64, affine=affine_par)
        for i in self.bn1.parameters():
            i.requires_grad = False
        self.relu = nn.ReLU(inplace=True)
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1, ceil_mode=True)  # change

        self.layer1 = self._make_layer(block, 64, layers[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
        self.layer3 = self._make_layer(block, 256, layers[2], stride=1, dilation=2)
        self.layer4 = self._make_layer(block, 512, layers[3], stride=1, dilation=4)

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                m.weight.data.normal_(0, 0.01)
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_()

    def _make_layer(self, block, planes, blocks, stride=1, dilation=1):
        downsample = None
        if stride != 1 or self.inplanes != planes * block.expansion or dilation == 2 or dilation == 4:
            downsample = nn.Sequential(
                nn.Conv2d(self.inplanes, planes * block.expansion,
                          kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(planes * block.expansion, affine=affine_par))
        for i in downsample._modules['1'].parameters():
            i.requires_grad = False
        layers = []
        layers.append(block(self.inplanes, planes, stride, dilation=dilation, downsample=downsample))
        self.inplanes = planes * block.expansion
        for i in range(1, blocks):
            layers.append(block(self.inplanes, planes, dilation=dilation))
        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)
        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)
        return x


class Deeplab(nn.Module):
    def __init__(self, backbone, classifier, num_classes):
        super(Deeplab, self).__init__()
        self.backbone = backbone
        self.classifier = classifier
        self.num_classes = num_classes

    def forward(self, x):
        x = self.backbone(x)
        y = self.classifier(x)
        return y

    def get_1x_lr_params_NOscale(self):
        """
        This generator returns all the parameters of the net except for
        the last classification layer. Note that for each batchnorm layer,
        requires_grad is set to False in deeplab_resnet.py, therefore this function does not return
        any batchnorm parameter
        """
        layers = [self.backbone.conv1, self.backbone.bn1,
                self.backbone.layer1, self.backbone.layer2, self.backbone.layer3, self.backbone.layer4]
        for layer in layers:
            for module in layer.modules():
                for param in module.parameters():
                    if param.requires_grad:
                        yield param

    def get_10x_lr_params(self):
        """
        This generator returns all the parameters for the last layer of the net,
        which does the classification of pixel into classes
        """
        for param in self.classifier.parameters():
            yield param

    def get_parameters(self, lr=1.):
        return [
            {'params': self.get_1x_lr_params_NOscale(), 'lr': 0.1 * lr},
            {'params': self.get_10x_lr_params(), 'lr': lr}
        ]


[docs]def deeplabv2_resnet101(num_classes=19, pretrained_backbone=True): """Constructs a DeepLabV2 model with a ResNet-101 backbone. Args: num_classes (int, optional): number of classes. Default: 19 pretrained_backbone (bool, optional): If True, returns a model pre-trained on ImageNet. Default: True. """ backbone = ResNet(Bottleneck, [3, 4, 23, 3]) if pretrained_backbone: # download from Internet saved_state_dict = load_state_dict_from_url(model_urls['deeplabv2_resnet101'], map_location=lambda storage, loc: storage, file_name="deeplabv2_resnet101.pth") new_params = backbone.state_dict().copy() for i in saved_state_dict: i_parts = i.split('.') if not i_parts[1] == 'layer5': new_params['.'.join(i_parts[1:])] = saved_state_dict[i] backbone.load_state_dict(new_params) classifier = ASPP_V2(2048, [6, 12, 18, 24], [6, 12, 18, 24], num_classes) return Deeplab(backbone, classifier, num_classes)

Docs

Access comprehensive documentation for Transfer Learning Library

View Docs

Tutorials

Get started for Transfer Learning Library

Get Started

Paper List

Get started for transfer learning

View Resources