It is mentioned in the paper that the internal terms can be implemented using the automatic differentiation provided by pytorch, but obviously internal term in Redner is also implemented by manual derivation.
How to use automatic differentiation and CUDA parallel computing to realize automatic differentiation of internal terms?
I would appreciate it if anyone could answer.
It is mentioned in the paper that the internal terms can be implemented using the automatic differentiation provided by pytorch, but obviously internal term in Redner is also implemented by manual derivation. How to use automatic differentiation and CUDA parallel computing to realize automatic differentiation of internal terms? I would appreciate it if anyone could answer.