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

[main] [refactor] refactor common_fused_moe.py (#2706)

Browse Source 
### What this PR does / why we need it?
1. Move prepare/finalize operation from moe_comm_method to
/ops/moe/fused_moe_prepare_and_finalize
2. Adapt to token_dispatcher in moe_comm_method
3. Move
moe_comm_method/experts_selector/token_dispatcher/fused_moe_prepare_and_finalize
to /ops/moe
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
e2e & ut

- vLLM version: v0.10.1.1
- vLLM main:
f4962a6d55

Signed-off-by: weichen <calvin_zhu0210@outlook.com>
Signed-off-by: Pr0Wh1teGivee <calvin_zhu0210@outlook.com>
Co-authored-by: weijinqian0 <12153182+weijinqian0@users.noreply.github.com>
This commit is contained in:
weichen
2025-09-08 20:09:50 +08:00
committed by GitHub
parent 1a82b16355
commit a041d4f328
21 changed files with 1052 additions and 932 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm_ascend/quantization/w4a8_dynamic.py
View File

@@ -27,7 +27,7 @@ from vllm.forward_context import get_forward_context
from vllm_ascend.ascend_forward_context import FusedMoEState
from vllm_ascend.distributed.parallel_state import get_mc2_group
from vllm_ascend.ops.fused_moe import unified_fused_experts_eager
from vllm_ascend.ops.layers.experts_selector import select_experts
from vllm_ascend.ops.moe.experts_selector import select_experts
class AscendW4A8DynamicLinearMethod: