I’m following the finetuning tutorial for inception_v3 (all the other models work), but I have the following error: ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1, 768, 1, 1]).
The code is currently the exact same as the tutorial, is this outdated or should I change something?
Most likely some batchnorm layer is throwing this error, since you are only using a single sample in the batch containing only one pixel.
Try to increase the batch size and run it again.