ReResNet源码解读

程序运行顺序：ReResNet–ResLayer–Bottleneck

class ReResNet(BaseBackbone)
class ResLayer(nn.Sequential)
class Bottleneck(enn.EquivariantModule)

以运行train.py为例，按照程序运行流程解读。

python tools/train.py /configs/re_resnet/re_resnet50_c8_batch256.py

1. List item

_in_channels = stem_channels
_out_channels = base_channels * self.expansion
for i, num_blocks in enumerate(self.stage_blocks):
    stride = strides[i]
    dilation = dilations[i]
    res_layer = self.make_res_layer(
        block=self.block,
        num_blocks=num_blocks,
        in_channels=_in_channels,
        out_channels=_out_channels,
        expansion=self.expansion,
        stride=stride,
        dilation=dilation,
        style=self.style,
        avg_down=self.avg_down,
        with_cp=with_cp,
        conv_cfg=conv_cfg,
        norm_cfg=norm_cfg)
    _in_channels = _out_channels
    _out_channels *= 2
    layer_name = f'layer{i + 1}'
    self.add_module(layer_name, res_layer)
    self.res_layers.append(layer_name)

第一轮循环，_in_channels =64，_out_channels =256，self.stage_blocks=（3,4,6,3）。
self.make_res_layer调用ResLayer，开始构建神经网络。

2. ResLayer
   首先判断是否下采样(downsample)，对应Resnet论文中提到的短连接。

        layers = []
        layers.append(
            block(
                in_channels=in_channels,
                out_channels=out_channels,
                expansion=self.expansion,
                stride=stride,
                downsample=downsample,
                conv_cfg=conv_cfg,
                norm_cfg=norm_cfg,
                **kwargs))
        in_channels = out_channels

开始构建第一层网络，调用Bottleneck

3. Bottleneck

        self.norm1_name, norm1 = build_norm_layer(
            norm_cfg, self.mid_channels, postfix=1)
        self.norm2_name, norm2 = build_norm_layer(
            norm_cfg, self.mid_channels, postfix=2)
        self.norm3_name, norm3 = build_norm_layer(
            norm_cfg, out_channels, postfix=3)

构建BN层

        self.conv1 = build_conv_layer(
            conv_cfg,
            in_channels,
            self.mid_channels,
            kernel_size=1,
            stride=self.conv1_stride,
            bias=False)
        self.add_module(self.norm1_name, norm1)
        self.relu1 = ennReLU(self.mid_channels)
        self.conv2 = build_conv_layer(
            conv_cfg,
            self.mid_channels,
            self.mid_channels,
            kernel_size=3,
            stride=self.conv2_stride,
            padding=dilation,
            dilation=dilation,
            bias=False)
        self.add_module(self.norm2_name, norm2)
        self.relu2 = ennReLU(self.mid_channels)
        self.conv3 = build_conv_layer(
            conv_cfg,
            self.mid_channels,
            out_channels,
            kernel_size=1,
            bias=False)
        self.add_module(self.norm3_name, norm3)
        self.relu3 = ennReLU(out_channels)

构建卷积层，一共三层卷积，每一层都是由卷积+BN+Relu构成。

1. ResLayer
   卷积层构建完后，append到layers里面，开始生成剩下num_blocks-1个神经网络，这里对应Resnet论文中残差提取特征。然后调用Bottleneck生成卷积层。

        for i in range(1, num_blocks):
            layers.append(
                block(
                    in_channels=in_channels,
                    out_channels=out_channels,
                    expansion=self.expansion,
                    stride=1,
                    conv_cfg=conv_cfg,
                    norm_cfg=norm_cfg,
                    **kwargs))

5. 网络全貌
   
   生成4个layer

layer3是6，对应self.stage_blocks=（3,4,6,3）

layer3的第一层，包含三个卷积三个BN，一个下采样层。