Open
Conversation
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The original generate.py file set the whole (b, s, v) logit tensor to 64 bit for adding gumbel noise. This PR adds a version which does it sequentially over the batch, saving up to a factor of batch_size VRAM. The time overhead is minimal. The file generate_optimized.py contains the new script, with test_efficient_gumbel.py testing the approach for validity and confirming the significant VRAM savings. Output from the test file is also included in gumbel_test_output.txt, showing the significant VRAM savings and minimal time overhead. It also identifies and fixes a possible bug when setting the eot/eos confidences to -inf.