pytorch -> onnx 报错：RuntimeError:

参考：https://github.com/pytorch/pytorch/issues/8392
原因：在用 pytorch pretrained resnet 模型时，下面红框中的赋值部分，其实是引用 resnet，而不是显示的 layer 定义。

解决办法

在定义模型时，指定每一层 layer 的显示定义，得到模型之后，再把 pretrained resnet 的参数赋给定义好的模型的每一层。

以重载 resnet18 为例

• 原始模型定义，__init__ 中并没有显示指定每个 layer

import torch.nn as nn
from torchvision import models


class resnet18(nn.Module):
    def __init__(self, pretrained=True):
        super().__init__()
        self.features = models.resnet18(pretrained=pretrained)
        self.conv1 = self.features.conv1
        # self.conv1 = nn.Conv2d(4, 64, kernel_size=7, stride=2, padding=3, bias=False)  # 4 channels
        self.bn1 = self.features.bn1
        self.relu = self.features.relu
        self.maxpool1 = self.features.maxpool
        self.layer1 = self.features.layer1
        self.layer2 = self.features.layer2
        self.layer3 = self.features.layer3
        self.layer4 = self.features.layer4
        # GAP
        self.gap = nn.AdaptiveAvgPool2d(output_size=(1, 1))

    def forward(self, input):
        x = self.conv1(input)
        x = self.relu(self.bn1(x))
        x = self.maxpool1(x)
        feature1 = self.layer1(x)  # 1 / 4
        feature2 = self.layer2(feature1)  # 1 / 8
        feature3 = self.layer3(feature2)  # 1 / 16
        feature4 = self.layer4(feature3)  # 1 / 32
        tail = self.gap(feature4)
        return feature3, feature4, tail

• 修改后，先 copy 默认的 ResNet，然后在函数 resnet18 中进行参数赋值

class ResNet(nn.Module):

    def __init__(self, block, layers, num_classes=1000, zero_init_residual=False):
        super(ResNet, self).__init__()
        self.inplanes = 64
        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)
        # GAP
        self.gap = nn.AdaptiveAvgPool2d(output_size=(1, 1))

        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.init.constant_(m.weight, 1)
                nn.init.constant_(m.bias, 0)

        # 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):
        downsample = None
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = nn.Sequential(
                conv1x1(self.inplanes, planes * block.expansion, stride),
                nn.BatchNorm2d(planes * block.expansion),
            )

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

        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        feature1 = self.layer1(x)  # 1 / 4
        feature2 = self.layer2(feature1)  # 1 / 8
        feature3 = self.layer3(feature2)  # 1 / 16
        feature4 = self.layer4(feature3)  # 1 / 32
        tail = self.gap(feature4)
        return feature3, feature4, tail


def resnet18(pretrained=False, **kwargs):
    """Constructs a ResNet-18 model.

    Args:
        pretrained (bool): If True, returns a model pre-trained on ImageNet
    """
    model = ResNet(BasicBlock, [2, 2, 2, 2], **kwargs)
    if pretrained: # 在这里 copy pretrained resnet 参数
        features = models.resnet18(pretrained=pretrained)
        model.conv1 = features.conv1
        model.bn1 = features.bn1
        model.relu = features.relu
        model.maxpool = features.maxpool
        model.layer1 = features.layer1
        model.layer2 = features.layer2
        model.layer3 = features.layer3
        model.layer4 = features.layer4
    return model