This repository has been archived by the owner on Dec 16, 2022. It is now read-only.
Several micro optimizations #4833
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
epwalsh
commented
Dec 2, 2020
| @@ -1548,7 +1548,6 @@ def add_sentence_boundary_token_ids( | |||
| The new mask for the tensor, taking into account the appended tokens | |||
| marking the beginning and end of the sentence. | |||
| """ | |||
| # TODO: matthewp, profile this transfer | |||
There was a problem hiding this comment.
I benchmarked this. This function is actually faster this way than with keeping sequence_lengths on GPU with:
sequence_lengths = mask.sum(dim=1).detach()
AkshitaB
approved these changes
Dec 2, 2020
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Mainly replacing calls where a tensor is created and then sent to a device with
.to(device)with calls that create the tensor directly on the device, which is over twice as fast. You can run the benchmarks yourself with:pytest -c benchmarks/pytest.ini benchmarks/nn/util_bench.py -k 'create_tensor'These are the results I got: