Input type (CUDAFloatTensor) and weight type (CPUFloatTensor) should be the same

1

When I run my model, it returns me Input type (CUDAFloatTensor) and weight type (CPUFloatTensor) should be the same
I think that’s because I used a pretrained model and used the weights of that pretrained model maybe, because the error returns from there.
Any idea to fix it?

1 Like
2

In PyTorch, you need to make sure that the operands are in the same context device(both on CPU or both on GPU). Try to convey the model into GPU using:

model = model.cuda()
1 Like
3

and then it returns meRuntimeError: cuda runtime error (2) : out of memory at /pytorch/torch/lib/THC/generic/THCStorage.cu:58