我的 PyTorch 转发函数可以做额外的操作吗?

Can my PyTorch forward function do additional operations?

通常一个 forward 函数将一堆层和 returns 最后一层的输出串在一起。在返回之前,我可以在最后一层之后做一些额外的处理吗?例如,一些标量乘法和整形通过 .view?

我知道 autograd 以某种方式计算出梯度。所以我不知道我的额外处理是否会以某种方式搞砸了。谢谢。

tracks the gradients via the computational graph of the tensors, not through the functions. As long as your tensors has requires_grad=True property and their grad 不是 None 你可以(几乎)做任何你喜欢的事并且仍然可以反向传播。
只要您使用的是 pytorch 的操作(例如 here and here 中列出的操作),您就应该没问题。

有关详细信息,请参阅 this

例如(取自torchvision's VGG implementation):

class VGG(nn.Module):

    def __init__(self, features, num_classes=1000, init_weights=True):
        super(VGG, self).__init__()
        #  ...

    def forward(self, x):
        x = self.features(x)
        x = self.avgpool(x)
        x = torch.flatten(x, 1)  # <-- what you were asking about
        x = self.classifier(x)
        return x

更复杂的例子见torchvision's implementation of ResNet:

class Bottleneck(nn.Module):
    def __init__(self, inplanes, planes, stride=1, downsample=None, groups=1,
                 base_width=64, dilation=1, norm_layer=None):
        super(Bottleneck, self).__init__()
        # ...

    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)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        if self.downsample is not None:    # <-- conditional execution!
            identity = self.downsample(x)

        out += identity  # <-- inplace operations
        out = self.relu(out)

        return out