modeling_resnet.py

import math

import torch.nn as nn
from transformers import PreTrainedModel

from .configuration_resnet import ResEncoderConfig


def conv3x3(in_planes, out_planes, stride=1):
    return nn.Conv2d(
        in_planes, out_planes, kernel_size=3, stride=stride, padding=1, bias=False
    )


def downsample_basic_block(inplanes, outplanes, stride):
    return nn.Sequential(
        nn.Conv2d(inplanes, outplanes, kernel_size=1, stride=stride, bias=False),
        nn.BatchNorm2d(outplanes),
    )


def downsample_basic_block_v2(inplanes, outplanes, stride):
    return nn.Sequential(
        nn.AvgPool2d(
            kernel_size=stride, stride=stride, ceil_mode=True, count_include_pad=False
        ),
        nn.Conv2d(inplanes, outplanes, kernel_size=1, stride=1, bias=False),
        nn.BatchNorm2d(outplanes),
    )


class BasicBlock(nn.Module):
    expansion = 1

    def __init__(self, inplanes, planes, stride=1, downsample=None, relu_type="relu"):
        super(BasicBlock, self).__init__()

        assert relu_type in ["relu", "prelu"]

        self.conv1 = conv3x3(inplanes, planes, stride)
        self.bn1 = nn.BatchNorm2d(planes)

        if relu_type == "relu":
            self.relu1 = nn.ReLU(inplace=True)
            self.relu2 = nn.ReLU(inplace=True)
        elif relu_type == "prelu":
            self.relu1 = nn.PReLU(num_parameters=planes)
            self.relu2 = nn.PReLU(num_parameters=planes)
        else:
            raise Exception("relu type not implemented")

        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.relu1(out)
        out = self.conv2(out)
        out = self.bn2(out)
        if self.downsample is not None:
            residual = self.downsample(x)

        out += residual
        out = self.relu2(out)

        return out


class ResNet(nn.Module):
    def __init__(
        self,
        block,
        layers,
        num_classes=1000,
        relu_type="relu",
        gamma_zero=False,
        avg_pool_downsample=False,
    ):
        self.inplanes = 64
        self.relu_type = relu_type
        self.gamma_zero = gamma_zero
        self.downsample_block = (
            downsample_basic_block_v2 if avg_pool_downsample else downsample_basic_block
        )

        super(ResNet, self).__init__()
        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.AdaptiveAvgPool2d(1)

        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(2.0 / n))
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_()

        if self.gamma_zero:
            for m in self.modules():
                if isinstance(m, BasicBlock):
                    m.bn2.weight.data.zero_()

    def _make_layer(self, block, planes, blocks, stride=1):
        downsample = None
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = self.downsample_block(
                inplanes=self.inplanes,
                outplanes=planes * block.expansion,
                stride=stride,
            )

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

        return nn.Sequential(*layers)

    def forward(self, 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)
        return x


class ResEncoder(PreTrainedModel):
    def __init__(self, config: ResEncoderConfig):
        super(ResEncoder, self).__init__(config=config)
        self.frontend_nout = config.frontend_nout
        self.backend_out = config.backend_out
        frontend_relu = (
            nn.PReLU(num_parameters=self.frontend_nout)
            if config.relu_type == "prelu"
            else nn.ReLU()
        )
        self.frontend3D = nn.Sequential(
            nn.Conv3d(
                1,
                self.frontend_nout,
                kernel_size=(5, 7, 7),
                stride=(1, 2, 2),
                padding=(2, 3, 3),
                bias=False,
            ),
            nn.BatchNorm3d(self.frontend_nout),
            frontend_relu,
            nn.MaxPool3d(kernel_size=(1, 3, 3), stride=(1, 2, 2), padding=(0, 1, 1)),
        )
        self.trunk = ResNet(BasicBlock, [2, 2, 2, 2], relu_type=config.relu_type)

    def forward(self, x):
        B, C, T, H, W = x.size()
        x = self.frontend3D(x)
        Tnew = x.shape[2]
        x = self.threeD_to_2D_tensor(x)
        x = self.trunk(x)
        x = x.view(B, Tnew, x.size(1))
        x = x.transpose(1, 2).contiguous()
        return x

    def threeD_to_2D_tensor(self, x):
        n_batch, n_channels, s_time, sx, sy = x.shape
        x = x.transpose(1, 2).contiguous()
        return x.reshape(n_batch * s_time, n_channels, sx, sy)