If The brand new shape ends in more elements than the first tensor, new elements are going to be uninitialized in memory. You need to flatten this to provide it on the fully related layer. So that you convey to PyTorch to reshape the tensor you obtained to have particular https://thca-good-health-benefits78898.total-blog.com/the-smart-trick-of-industry-experts-that-nobody-is-discussing-64914564