Fix Pytorch models

In this MR, some bugs related to PyTorch models are fixed and improved:

1. add to(device),
2. add model.eval(),
3. add torch.no_grad()