xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-02 12:44:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

[gpt-oss] Fix mxfp4 support (#22700)

Browse Source 
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
This commit is contained in:
Chen Zhang 2025-08-11 21:22:26 -07:00 committed by GitHub
parent 4678503476
commit bbaf9e9cb1
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/layers/quantization/utils/mxfp4_utils.py
View File

@ -68,7 +68,7 @@ def _can_support_mxfp4(use_grouped_topk: bool = False,
return not (use_grouped_topk or topk_group or num_expert_group return not (use_grouped_topk or topk_group or num_expert_group
or expert_map or custom_routing_function or expert_map or custom_routing_function
or e_score_correction_bias or apply_router_weight_on_input or e_score_correction_bias or apply_router_weight_on_input
or scoring_func != "softmax" or activation != "silu" or scoring_func != "softmax" or activation != "swiglu_oai"
or expert_load_view or logical_to_physical_map or expert_load_view or logical_to_physical_map
or logical_replica_count) or logical_replica_count)