Pass packed boundary metadata to Qwen3.5 linear-attention fast kernels from data collator#45034
Draft
sdharani91 wants to merge 8 commits intohuggingface:mainfrom
Draft
Pass packed boundary metadata to Qwen3.5 linear-attention fast kernels from data collator#45034sdharani91 wants to merge 8 commits intohuggingface:mainfrom
sdharani91 wants to merge 8 commits intohuggingface:mainfrom
Conversation
…s for issue 44717
Contributor
|
[For maintainers] Suggested jobs to run (before merge) run-slow: qwen3_5 |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
This is a follow up to #44867
This PR fixes Qwen3.5 padding-free packed inputs on the linear-attention fast path by consuming collator-provided packed metadata. The linear-attention block now uses seq_idx for the causal convolution path and cu_seq_lens_q / cu_seq_lens_k for the FLA path, matching the repo’s existing DataCollatorWithFlattening contract. I also added a deterministic fast-path regression test comparing padded and padding-free inputs, plus a slow-path contract test that raises clearly when padding-free kwargs are passed without fast-kernel support. The slow fallback implementation itself is unchanged in this PR.
Fixes # 44717
Code Agent Policy
The Transformers repo is currently being overwhelmed by a large number of PRs and issue comments written by
code agents. We are currently bottlenecked by our ability to review and respond to them. As a result,
we ask that new users do not submit pure code agent PRs at this time.
You may use code agents in drafting or to help you diagnose issues. We'd also ask autonomous "OpenClaw"-like agents
not to open any PRs or issues for the moment.
PRs that appear to be fully agent-written will probably be closed without review, and we may block users who do this
repeatedly or maliciously.
This is a rapidly-evolving situation that's causing significant shockwaves in the open-source community. As a result,
this policy is likely to be updated regularly in the near future. For more information, please read
CONTRIBUTING.md.Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
@vasqu