You must flatten this to present it on the fully linked layer. So you inform PyTorch to reshape the tensor you acquired to get unique amount of columns and convey to it to decide the amount of rows by by itself. Read story Dust Firyal Quraishi Winner of Contest #325 https://carcrashcompensation94714.jts-blog.com/38263507/not-known-factual-statements-about-trusted-partner