Vladimir Iglovikov

:+1:

:+1:

I confirm. When I use 2 GPUs and call resnet50 from torchvision => works well. Call it from Cadene pre-trained models => `RuntimeError: Expected tensor for argument #1 'input' to...

Wrapper like: ``` class Net(nn.Module): def __init__(self, model): super(Net, self).__init__() self.l1 = nn.Sequential(*list(model.children())[:-1]).to('cuda:0') self.last = list(model.children())[-1] def forward(self, x): x = self.l1(x) x = x.view(x.size()[0], -1) x = self.last(x) return...

@Cadene

I loved all the issues when moved to https://github.com/rwightman/pytorch-image-models

First of all `num train = 0, num_val = 0` looks strange. Are you sure that your DataLoader defined in https://github.com/ternaus/robot-surgery-segmentation/blob/master/dataset.py is correct?

Second `model.load_state_dict(state['model'])` is trying to load a model which is happening when your folder `runs/debug` is not empty. Can you delete it and try again?

I will look at it after the weekend, but if you know how to implement caching - feel free to create PR, it would be great and may help other...

We are working on a document with this information, but it is not ready yet.