Does creating a new tensor break Autograd?

[AlphaBetaGamma96]

Hi,

I think using torch.tensor would break it because torch.tensor has no grad_fn. What you could do is replace,

@deepfailure

new_output = torch.tensor([tensor.sum() for tensor in torch.split(output, 10, dim=1)])

with,

new_output = torch.stack([tensor.sum() for tensor in torch.split(output, 10, dim=1)])

That should keep the gradients flowing as torch.stack has a grad_fn whereas torch.tensor doesn’t!