tensorplay.distributed.autograd.backward

tensorplay.distributed.autograd.backward(context_id: int, roots: Any, retain_graph: bool = False) → None[source]