Skip to content

Fix precision issue - #33

Open
WangLedi wants to merge 7 commits into
ethan686:dit_refactor_tp_quantfrom
WangLedi:dit_refactor_tp_quant
Open

Fix precision issue #33
WangLedi wants to merge 7 commits into
ethan686:dit_refactor_tp_quantfrom
WangLedi:dit_refactor_tp_quant

Conversation

@WangLedi

Copy link
Copy Markdown

Description

Related Issues

Change Type

  • Bug fix
  • New feature
  • Performance improvement
  • Refactor
  • Documentation
  • Test
  • Build or CI

Pull Request Checklist

Thank you for contributing to xLLM. Before requesting review, please make sure the following items are complete.

PR Title and Commit Messages

  • The PR title and each commit message follow the xLLM commit format: <type>: <subject>.

Allowed types: feat, bugfix, docs, test, refactor, chore, style, revert, perf, model, build, release.
The subject should use clear English, start with a verb, include at least 4 words, and end with ..

Pre-commit Checks

  • I have installed pre-commit by running pip install pre-commit or an equivalent command.
  • I have installed the hooks with pre-commit install.
  • I have run pre-commit run --all-files and fixed any reported issues.

If you are unsure how to set up pre-commit, see the pre-commit documentation.

Self Review

  • I have self-reviewed the code according to .agents/skills/code-review/references/custom-code-style.md, especially code written or assisted by AI.
  • I have rebased this PR onto the latest main branch.

Build and Test Coverage

  • Tests have been added or updated as needed.
  • CUDA: python setup.py build test has passed on a CUDA machine.
  • NPU: python setup.py build test has passed on an NPU machine.
  • MLU: python setup.py build test has passed on an MLU machine.

Reviewer Notes

Super User and others added 7 commits July 23, 2026 10:49
* Quantized Weight Layer Name Conversion

* Quantized Weight Layer Name Conversion
…llelLinear in Wan2.2 TP

- Replace WanGELU DiTParallelLinear with ColumnParallelLinear
- Replace WanFeedForward DiTParallelLinear with RowParallelLinear
- Rewrite WanAttention members and constructor with ColumnParallelLinear/RowParallelLinear
- Extract SP all2all from linear layers to explicit calls in WanAttention::forward()
- Update WanAttention verify_loaded_weights for new linear types
- Create single-rank process groups for tp_size=1 and remove fallback
- Fix dit_collective_communicator TP context type
@ethan686
ethan686 force-pushed the dit_refactor_tp_quant branch from e8446fc to 224388e Compare July 28, 2026 01:11
@WangLedi
WangLedi force-pushed the dit_refactor_tp_quant branch from ca77a6b to d947f76 Compare July 28, 2026 01:38
@ethan686
ethan686 force-pushed the dit_refactor_tp_quant branch 2 times, most recently from 3c7ed0b to ae4fce0 Compare July 29, 2026 01:33
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