Closed ffuuugor closed 1 year ago
Since PyTorch 1.13 was released, our runtime/memory benchmarks are failing on the recurrent modules (GRU/RNN/LSTM)
The reason is - our implementation did not change, but the pytorch baseline got faster - by roughly the factor of 3.
Changing the thresholds based on the latest performance on CircleCI machine (+15% margin)
@ffuuugor has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.
Since PyTorch 1.13 was released, our runtime/memory benchmarks are failing on the recurrent modules (GRU/RNN/LSTM)
The reason is - our implementation did not change, but the pytorch baseline got faster - by roughly the factor of 3.
Changing the thresholds based on the latest performance on CircleCI machine (+15% margin)