transformers
Add expert parallelism (EP) config support for Qwen3 MoE
#45436
Merged

Add expert parallelism (EP) config support for Qwen3 MoE #45436

AmineDiro
HuggingFaceDocBuilderDev
AmineDiro
3outeille
AmineDiro AmineDiro changed the title Add expert parallelism (EP) support for Qwen3 MoE + fix GroupedGemmParallel for 2D meshes Add expert parallelism (EP) support for Qwen3 MoE 98 days ago
AmineDiro AmineDiro changed the title Add expert parallelism (EP) support for Qwen3 MoE Add expert parallelism (EP) config support for Qwen3 MoE 98 days ago
ArthurZucker
ArthurZucker approved these changes on 2026-04-21
ArthurZucker
AmineDiro
AmineDiro Add base_model_ep_plan to Qwen3OmniMoeTalkerTextConfig
d168c968
AmineDiro Add expert-only EP plan to Qwen3 MoE configs
1b4b8828
AmineDiro Use expert-only EP plan for Qwen3 MoE configs
dbf68c04
AmineDiro Remove comments from generated qwen3_vl_moe config
a29cc422
AmineDiro AmineDiro force pushed from 875002b9 to a29cc422 97 days ago
AmineDiro Merge branch 'main' into qwen3-moe-ep-v2
38181267
ArthurZucker
ArthurZucker
ArthurZucker
github-actions
github-actions
AmineDiro
ArthurZucker
ArthurZucker
github-actions
github-actions[bot] Apply repo consistency fixes
907b79d8
AmineDiro Merge branch 'main' into qwen3-moe-ep-v2
eb992ed7
AmineDiro Add base_model_ep_plan to qwen3_vl_moe and qwen3_omni_moe modular fil…
abdac348
AmineDiro Merge branch 'main' into qwen3-moe-ep-v2
cbe7a028
AmineDiro Fix base_model_ep_plan position in generated configs to match modular…
7276c5a8
github-actions
AmineDiro
AmineDiro Merge branch 'main' into qwen3-moe-ep-v2
a8d6232f
ArthurZucker ArthurZucker merged 7cf4241f into main 95 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone