Skip to content

Conversation

@Sunny-bot1
Copy link
Collaborator

  1. 由于 machete weight only gemm 的权重量化和存储方式不同,本PR为其适配了v1 loader;
  2. 在单测 test_common_model.py 中为 DeepSeek 的 case 开启 FD_USE_MACHETE ,增加覆盖率;
  3. 在 Qwen3-30B-A3B wint4 tp1/2 已验证逐token对齐。

@paddle-bot
Copy link

paddle-bot bot commented Sep 9, 2025

Thanks for your contribution!

@zhoutianzi666 zhoutianzi666 merged commit 3b1da6e into PaddlePaddle:develop Sep 10, 2025
45 of 50 checks passed
Sunny-bot1 added a commit to Sunny-bot1/FastDeploy that referenced this pull request Sep 18, 2025
Jiang-Jia-Jun pushed a commit that referenced this pull request Sep 19, 2025
* support v1 loader for machete (#3999)

* [Optimize] Support WINT8 and group scale for Machete (#3905)

* [Optimize] Machete using group scale default (#4121)
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.

4 participants