unstack() in tensorflow is the inverse of stack() (the latter of which is also available in pytorch). I can imagine that there is a function in pytorch, that takes for instance a MxN tensor and returns a list of M N-dimensional tensors (the rows) or N M-dimensional tensors (the columns), but I couldn’t find it.
TF documentation of this function is here:
https://www.tensorflow.org/api_docs/python/tf/unstack
Is there a similar function in pytorch?