Zhai Yujie

Results 2 issues of Zhai Yujie

- code: > state_dict = torch.hub.load_state_dict_from_url(checkpoint, progress=True) > if 'mask_values' in state_dict: state_dict.pop('mask_values') > net.load_state_dict(state_dict) Generally, ‘state_dict’ contains module names & parameters, but not mask values. What is the purpose...

For binary classification problems, I think n_classes should be 2 corresponding to mask_value={0, 1}. But when to use n_classes=1?