net.py

import torch
import torch.nn as nn
from torch.utils.model_zoo import load_url as load_state_dict_from_url
import context_module

model_urls = {
    '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, module=""):
        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

        if module:
            print("not implemented yet")
            raise ValueError

    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, opt, stride=1, downsample=None, groups=1,
                 base_width=64, dilation=1, norm_layer=None, module=""):
        super(Bottleneck, self).__init__()
        if norm_layer is None:
            norm_layer = nn.BatchNorm2d
        width = int(planes * (base_width / 64.)) * groups

        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
        self.module_type = module

        if module == "none":
            self.module = None
        elif module == 'acm':
            self.module = context_module.ACM(num_heads=opt.num_acm_groups, num_features=planes * 4)
            self.module.init_parameters()
        else:
            raise ValueError("undefined module")

    def forward(self, x):

        if isinstance(x, tuple):
            x, prev_dp = x
        else:
            prev_dp = None

        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)

        dp = None
        if self.module is not None:
            out = self.module(out)
            if isinstance(out, tuple):
                out, dp = out
                if prev_dp is not None:
                    dp = prev_dp + dp

        out += identity
        out = self.relu(out)

        if dp is None:
            return out
        else:
            # diff loss
            return out, dp


class ResNet(nn.Module):

    def __init__(self, block, layers, opt, num_classes=1000, zero_init_residual=False,
                 groups=1, width_per_group=64, replace_stride_with_dilation=None,
                 norm_layer=None, module=""):
        super(ResNet, self).__init__()
        self.opt = opt
        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, 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)

        assert module in ['none', 'acm']
        self.layer1 = self._make_layer(block, 64,  layers[0], 1, dilate=False,                           module=module)
        self.layer2 = self._make_layer(block, 128, layers[1], 2, dilate=replace_stride_with_dilation[0], module=module)
        self.layer3 = self._make_layer(block, 256, layers[2], 2, dilate=replace_stride_with_dilation[1], module=module)
        self.layer4 = self._make_layer(block, 512, layers[3], 2, dilate=replace_stride_with_dilation[2], module=module)
        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
        self.fc = nn.Linear(512 * block.expansion, num_classes)

        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.InstanceNorm2d, nn.LayerNorm)):
                try:
                    nn.init.constant_(m.weight, 1)
                    nn.init.constant_(m.bias, 0)
                except:
                    print("Module without affine doesnt have weights")

        # 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, module=""):
        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 = []

        # configure module placement
        module_placement = [module] * blocks
        layers.append(block(self.inplanes, planes, self.opt, stride, downsample, self.groups,
                            self.base_width, previous_dilation, norm_layer, module=module_placement[0]))
        self.inplanes = planes * block.expansion
        for block_idx in range(1, blocks):

            layers.append(block(self.inplanes, planes, self.opt, groups=self.groups,
                                base_width=self.base_width, dilation=self.dilation,
                                norm_layer=norm_layer, module=module_placement[block_idx]))

        return nn.Sequential(*layers)

    def _forward_body(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

    def forward(self, x):
        x = self._forward_body(x)
        if isinstance(x, tuple):
            x, dp = x
        else:
            dp = None

        y = self._forward_task(self, x)

        if dp is not None:
            return y, dp
        else:
            return y


def _resnet(arch, block, layers, pretrained, progress, **kwargs):
    model = ResNet(block, layers, **kwargs)
    if pretrained:
        state_dict = load_state_dict_from_url(model_urls[arch],
                                              progress=progress)
        try:
            model.load_state_dict(state_dict)
        except:
            print("keys did not mathc")
            print("entering custom loading statedict")
            existing_statedict = model.state_dict()
            cooccuring = {}
            for key, val in state_dict.items():
                if key in existing_statedict:
                    cooccuring[key] = state_dict[key]
                else:
                    print(key, "does not exists in the new model")
            print("keys adding ", len(cooccuring.keys()))
            print("whole keys", len(existing_statedict.keys()))
            existing_statedict.update(cooccuring)
            model.load_state_dict(existing_statedict)

    return model


def resnet34(pretrained=False, progress=True, **kwargs):
    r"""ResNet-34 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('resnet34', BasicBlock, [3, 4, 6, 3], pretrained, progress,
                   **kwargs)


def resnet50(pretrained=False, progress=True, **kwargs):
    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('resnet50', Bottleneck, [3, 4, 6, 3], pretrained, progress,
                   **kwargs)


def resnet101(pretrained=False, progress=True, **kwargs):
    r"""ResNet-101 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('resnet101', Bottleneck, [3, 4, 23, 3], pretrained, progress,
                   **kwargs)


def resnext50_32x4d(pretrained=False, progress=True, **kwargs):
    r"""ResNeXt-50 32x4d model from
    `"Aggregated Residual Transformation for Deep Neural Networks" <https://arxiv.org/pdf/1611.05431.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
    """
    kwargs['groups'] = 32
    kwargs['width_per_group'] = 4
    return _resnet('resnext50_32x4d', Bottleneck, [3, 4, 6, 3],
                   pretrained, progress, **kwargs)

def resnet(arch, pretrained, num_classes, zero_init_residual, module, opt):
    if arch == "resnet50":
        model = resnet50(pretrained=pretrained, progress=True, opt=opt, num_classes=num_classes,
                         zero_init_residual=zero_init_residual, module=module)
    elif arch == "resnet101":
        model = resnet101(pretrained=pretrained, progress=True, opt=opt, num_classes=num_classes,
                         zero_init_residual=zero_init_residual, module=module)
    elif arch == "resnext50_32x4d":
        model = resnet101(pretrained=pretrained, progress=True, opt=opt, num_classes=num_classes,
                         zero_init_residual=zero_init_residual, module=module)
    else:
        raise ValueError
    return model