I have a list of modules of size (Dim2), and an input of the shape Batch_size X Dim1 X Dim2. I want to pass the data through the list of modules where I want the module[i] to be applied to input[:,:,i]. Is there a clean way to do this? preferably using broadcasting.
Would this work:
outputs =  for m,i in zip(module,input.permute(2,0,1)): outputs.extend([m(i)])
broadcasting works for elementwise tensor arithmetic like add, multiply etc., i am not sure if it comes into play for passing inputs to functions.