The grad_input and grad_output are tuples that contain the gradientswith respect to the inputs and outputs respectively. The hook shouldnot modify its arguments, but it can optionally return a new gradient withrespect to the input that will be used in place of grad_input insubsequent computations. grad_input will only correspond to the inputs givenas positional arguments and all kwarg arguments are ignored. Entriesin grad_input and grad_output will be None for all non-Tensorarguments.