Implement Evolution Strategies on multiple processes or machines with PyTorch (Suggestions)

Hello! I am new to using PyTorch and would like some feedback as to the best way to implement an evolution strategies algorithm where I update the parameter of a model myself. Particularly, I would like a way to run the algorithm on multiple processes and possibly multiple machines and share data between them (return UpdatedModel parameters) so that I can update the parameters of my model myself. How would I do this in PyTorch. Please point me towards any useful documentation or suggestions. Thank you.