Skip to content

Latest commit

 

History

History
8 lines (7 loc) · 407 Bytes

cudnn_deterministic.rst

File metadata and controls

8 lines (7 loc) · 407 Bytes

Note

In some circumstances when using the CUDA backend with CuDNN, this operator may select a nondeterministic algorithm to increase performance. If this is undesirable, you can try to make the operation deterministic (potentially at a performance cost) by setting torch.backends.cudnn.deterministic = True. Please see the notes on :doc:`/notes/randomness` for background.