Skip to content

fix: handle Qwen3.5 in quantize_params_fp8 - #1093

Merged
Zhichenzzz merged 2 commits into
radixark:mainfrom
lawrence-harmonic:fix/quantize_qwen_3_5
May 8, 2026
Merged

fix: handle Qwen3.5 in quantize_params_fp8#1093
Zhichenzzz merged 2 commits into
radixark:mainfrom
lawrence-harmonic:fix/quantize_qwen_3_5

Conversation

@lawrence-harmonic

Copy link
Copy Markdown
Contributor

No description provided.

@lawrence-harmonic

lawrence-harmonic commented May 8, 2026

Copy link
Copy Markdown
Contributor Author

Copying from THUDM/slime#1683

@gemini-code-assist gemini-code-assist Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request updates the FP8 quantization processor to include linear attention parameters and improves the configuration handling for the quantization format. Feedback suggests replacing the use of assert with explicit ValueError exceptions for better production robustness and optimizing the parameter lookup process by moving the list of target weights to a module-level set.

@Zhichenzzz

Copy link
Copy Markdown
Contributor

LGTM! Thanks @lawrence-harmonic !

@Zhichenzzz
Zhichenzzz merged commit 323ccbd into radixark:main May 8, 2026
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.

2 participants