EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[bugfix] dep ineffective (#4417)

Browse Source 
### What this PR does / why we need it?
The expert mapping table and weights of the dynamic EPLB were not
updated, causing the accuracy to be correct but not effective. This bug
has now been fixed.

- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2

---------

Signed-off-by: shenchuxiaofugui <1311027364@qq.com>
This commit is contained in:
LI SHENGYONG
2025-11-29 15:18:29 +08:00
committed by GitHub
parent 8ebbf13c1a
commit 0151022ab8
3 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm_ascend/ops/fused_moe/fused_moe.py
View File

@@ -275,7 +275,7 @@ class AscendFusedMoE(FusedMoE):
return self.expert_map
def get_log2phy_map(self):
return self.logical_to_physical_map
return self.log2phy
def clear_moe_load(self):
if self.moe_load is not None: