Use of torch.manual_seed still results in different initial weights each time

8

I would say, note down the sizes during the for loop and after the for loop, create a clean model by setting the seed.

or you can reset the parameters at the end: