resnext.py

import math
import torch
import torch.nn as nn

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 BasicBlock(nn.Module):
    expansion = 1

    def __init__(self, inplanes, planes, stride=1, downsample=None):
        super(BasicBlock, self).__init__()
        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)
        self.relu = nn.ReLU(inplace=True)
        self.conv2 = conv3x3(planes, planes)
        self.bn2 = nn.BatchNorm2d(planes)
        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)

        if self.downsample is not None:
            residual = self.downsample(x)

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

        return out


class Bottleneck(nn.Module):
    expansion = 4

    def __init__(self, inplanes, planes, stride=1, downsample=None,
                 groups=1, base_width=64, is_last=False):
        super(Bottleneck, self).__init__()
        width = int(planes * (base_width / 64.)) * groups
        self.conv1 = nn.Conv2d(inplanes, width, kernel_size=1, bias=False)
        self.bn1 = nn.BatchNorm2d(width)
        self.conv2 = nn.Conv2d(width, width, kernel_size=3, stride=stride,
                               groups=groups, padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(width)
        self.conv3 = nn.Conv2d(width, planes * 4, kernel_size=1, bias=False)
        self.bn3 = nn.BatchNorm2d(planes * 4)
        self.relu = nn.ReLU(inplace=True)
        self.downsample = downsample
        self.stride = stride
        self.is_last = is_last

    def forward(self, x, relu=True):
        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
        if relu:
            out = self.relu(out)

        return out


class ResNext(nn.Module):

    def __init__(self, cfg, block, layers, groups=1, width_per_group=64, use_fc=False, dropout=None):
        self.inplanes = 64
        super(ResNext, self).__init__()
        self.cfg = cfg

        self.groups = groups
        self.base_width = width_per_group

        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
                               bias=False)
        self.bn1 = nn.BatchNorm2d(64)
        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)
        self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
        self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
        self.avgpool = nn.AvgPool2d(7, stride=1)

        self.use_fc = use_fc
        self.use_dropout = True if dropout else False

        if self.use_fc:
            print('Using fc.')
            self.fc_add = nn.Linear(512*block.expansion, 512)

        if self.use_dropout:
            print('Using dropout.')
            self.dropout = nn.Dropout(p=dropout)

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
                m.weight.data.normal_(0, math.sqrt(2. / n))
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_()

    def load_model(self, pretrain):
        print("Loading Backbone pretrain model from {}......".format(pretrain))
        model_dict = self.state_dict()
        pretrain_dict = torch.load(pretrain)
        pretrain_dict = pretrain_dict["state_dict"] if "state_dict" in pretrain_dict else pretrain_dict
        from collections import OrderedDict

        new_dict = OrderedDict()
        for k, v in pretrain_dict.items():
            if k.startswith("module"):
                k = k[7:]
            if "fc" not in k and "classifier" not in k:
                k = k.replace("encoder_q.", "")
                k = k.replace("backbone.", "")
                new_dict[k] = v

        model_dict.update(new_dict)
        self.load_state_dict(model_dict)
        print("Backbone model has been loaded......")

    def _make_layer(self, block, planes, blocks, stride=1, is_last=False):
        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),
                nn.BatchNorm2d(planes * block.expansion),
            )

        layers = []
        layers.append(block(self.inplanes, planes, stride, downsample,
                            groups=self.groups, base_width=self.base_width))
        self.inplanes = planes * block.expansion
        for i in range(1, blocks):
            layers.append(block(self.inplanes, planes,
                                groups=self.groups, base_width=self.base_width,
                                is_last=(is_last and i == blocks-1)))

        return nn.Sequential(*layers)

    def forward(self, x, *args):
        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 resnext50(
        cfg,
        pretrain=False,
        pretrained_model="",
        last_layer_stride=2,
):

    resnext = ResNext(cfg, Bottleneck, [3, 4, 6, 3], use_fc=False, dropout=None,
                      groups=32, width_per_group=4)

    if pretrain and pretrained_model != "":
        resnext.load_model(pretrain=pretrained_model)
    else:
        print("Choose to train from scratch")
    return resnext