Closed PhiSpel closed 3 months ago
when convert_to_tensor receives a tensor, torch.tensor(tensor, ...) is called. This raises an error, as the tensor is copied, which may be undesirable. In this case, torch.to() should be called with the device and dtype.
when convert_to_tensor receives a tensor, torch.tensor(tensor, ...) is called. This raises an error, as the tensor is copied, which may be undesirable. In this case, torch.to() should be called with the device and dtype.