Skip to content

[Ascend] support qwen35 mtp on Ascend-A3#4721

Open
wanfengcxz wants to merge 14 commits into
InternLM:mainfrom
wanfengcxz:wq/support_qwen35_mtp
Open

[Ascend] support qwen35 mtp on Ascend-A3#4721
wanfengcxz wants to merge 14 commits into
InternLM:mainfrom
wanfengcxz:wq/support_qwen35_mtp

Conversation

@wanfengcxz

@wanfengcxz wanfengcxz commented Jun 30, 2026

Copy link
Copy Markdown
Contributor

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily receiving feedbacks. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

support qwen35 MTP for ascend

Modification

Please briefly describe what modification is made in this PR.

BC-breaking (Optional)

Does the modification introduce changes that break the backward-compatibility of the downstream repositories?
If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.

Use cases (Optional)

If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.

Checklist

  1. Pre-commit or other linting tools are used to fix the potential lint issues.
  2. The modification is covered by complete unit tests. If not, please add more unit tests to ensure the correctness.
  3. If the modification has a dependency on downstream projects of a newer version, this PR should be tested with all supported versions of downstream projects.
  4. The documentation has been modified accordingly, like docstring or example tutorials.

tangzhiyi11 and others added 14 commits June 25, 2026 03:17
- op_backend.py: MTP detection (is_multi_token_decoding),
  effective_is_decoding, actual_seq_lengths_q, vendor_device_init trigger
- attention.py: add is_multi_token_decoding and actual_seq_lengths_q fields
- pagedattention.py: MTP verify reuses paged_prefill_attention
- config.py: SpecDecodeConfig.from_config add device_type param
- config_builder.py: pass device_type to SpecDecodeConfig

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Keep only the generic draft-step and accepted-token metadata plumbing in lmdeploy so the dlinfer backend can drive Ascend multi-token state updates without broad runtime hooks in the core runtime.

Made-with: Cursor
@wanfengcxz wanfengcxz changed the title Wq/support qwen35 mtp [Ascend] support qwen35 mtp on Ascend-A3 Jul 3, 2026
@wanfengcxz wanfengcxz marked this pull request as ready for review July 3, 2026 04:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants