I believe embeddings.numpy() on line 45 and 63 of saliency_interpreter.py should be embeddings.detach().cpu().numpy()
Otherwise while running the interpretation on GPU has a error 'TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.'
I believe embeddings.numpy() on line 45 and 63 of saliency_interpreter.py should be embeddings.detach().cpu().numpy() Otherwise while running the interpretation on GPU has a error 'TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.'