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

[MoE][TorchAir] Remove FusedMoEState (#4927)

Browse Source 
### What this PR does / why we need it?
Remove FusedMoEState which is used by torchair.

### Does this PR introduce _any_ user-facing change?
No

### How was this patch tested?
e2e & ut

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

Signed-off-by: weichen <calvin_zhu0210@outlook.com>
This commit is contained in:
weichen
2025-12-12 09:12:24 +08:00
committed by GitHub
parent bfafe30953
commit d54db76dd2
1 changed files with 0 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
vllm_ascend/ascend_forward_context.py
View File

@@ -19,15 +19,6 @@ else:
WeightPrefetchMethod = None WeightPrefetchMethod = None
class FusedMoEState(Enum):
AllGather = 0
All2All = 1
MC2 = 2
AllGatherEP = 3
NaiveMulticast = 4
All2AllSeq = 5
class MoECommType(Enum): class MoECommType(Enum):
ALLGATHER = 0 ALLGATHER = 0
MC2 = 1 MC2 = 1