I want to use resnet18, but I don’t need softmax. How do I get the output from the last conv layer?
The torchvision.models.resnet18
does not use a softmax as the last non-linearity, but uses an nn.Linear
layer to output the logits.
I want to use resnet18, but I don’t need softmax. How do I get the output from the last conv layer?
The torchvision.models.resnet18
does not use a softmax as the last non-linearity, but uses an nn.Linear
layer to output the logits.