Update stream llm to get correct outputs and re-enable rerotated-attention test. #656
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
monorimet
approved these changes
Apr 25, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
During the update of pytorch/HF, there seem to be a change of how causal mask was being handled. It seems like the attention.forward function used to get a
causal_maskfrom the argument asattention_maskwhen is_causal is on. Now it seems like we would need to construct our own mask whenis_causalis true. This was causing numerical issues in this test as well as on Llama2 qualitatively.This PR introduces construction of causal mask, as well as removing unnecessary tensor parallel config checks which simplifies the code quite a bit.