Skip to content

Conversation

@pathorn
Copy link
Contributor

@pathorn pathorn commented Feb 6, 2025

Finishes the rest of @LucasWilkinson's PR #12639

Implement chunked prefill for Triton MLA backend (DeepSeek v2/v3)

I have been testing with DeepSeek-Coder-v2-Lite-Instruct, both short and long-context responses seem correct.

My prompt is:
{"model": "deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct", "max_tokens": 100, "seed": 6, "messages": [{"role": "user", "content": "Yes or no, will the following code compile:\n\n.... some long python script ..."}

I think I fixed the issue with one block not being copied.

@github-actions
Copy link

github-actions bot commented Feb 6, 2025

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

@mergify
Copy link

mergify bot commented Feb 6, 2025

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @pathorn.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

@mergify mergify bot removed the needs-rebase label Feb 6, 2025
@pathorn pathorn changed the title Deep mla chunked Implement chunked prefill for Triton MLA attention backend Feb 6, 2025
@mergify
Copy link

mergify bot commented Feb 6, 2025

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @pathorn.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

@mergify mergify bot added the needs-rebase label Feb 6, 2025
@mergify mergify bot removed the needs-rebase label Feb 7, 2025
@pathorn pathorn force-pushed the deep-mla-chunked branch 2 times, most recently from 34d9cd8 to 858eb26 Compare February 7, 2025 06:43
@pathorn pathorn marked this pull request as ready for review February 7, 2025 06:50
LucasWilkinson and others added 2 commits February 11, 2025 00:10
Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Fix alignment and modulo error in remainder for _gather_kv_cache
Only use block_tables on chunked prefill
Allow enabling chunked prefill

Signed-off-by: Patrick Reiter Horn <patrick.horn@gmail.com>
Signed-off-by: Yessen Kanapin <yessen@deepinfra.com>
@mergify
Copy link

mergify bot commented Mar 5, 2025

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @pathorn.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

@mergify mergify bot added the needs-rebase label Mar 5, 2025
@LucasWilkinson
Copy link
Collaborator

superseded by: #12639

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants