You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fix a bug in FlashAttention when seqlen_q is not equal to seqlen_k (seqlen_v). For example, when KV-cache is enabled during inference, each step other than the first will have a query layer of sequence length=1.
The problem was covered in #302 and that one is already merged. Closing (please reopen if you believe that there is functionality here not included in the other PR).
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fix a bug in
FlashAttentionwhen seqlen_q is not equal to seqlen_k (seqlen_v). For example, when KV-cache is enabled during inference, each step other than the first will have a query layer of sequence length=1.The code is adapted from
FlashSelfAttentionof megatron-lm https://github.com/NVIDIA/Megatron-LM/blob/main/megatron/model/transformer.py#L375.