Bases: object
Gradient context manager, similar to torch.no_grad()
-
__enter__() → None[source]
Disables the computation graph while in the context
-
__exit__(exception_type, exception_value, exception_traceback) → None[source]
Reverts gradient option to what it was previously
