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

[Bugfix]remove mla patch, server args no need --compilation-config for ds v3.1 (#145)

Browse Source 
Signed-off-by: baoqian426 <1354987947@qq.com>
This commit is contained in:
baoqian426
2026-01-23 15:59:43 +08:00
committed by GitHub
parent 0ce5f1a3f7
commit 1eaa1336ac
2 changed files with 1 additions and 181 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm_kunlun/models/deepseek_v2.py
View File

@@ -54,7 +54,7 @@ from vllm.model_executor.layers.linear import (ColumnParallelLinear,
RowParallelLinear)
from vllm_kunlun.ops.linear import ReplicatedLinear
from vllm.model_executor.layers.logits_processor import LogitsProcessor
from vllm_kunlun.ops.attention.mla import MLAModules, MultiHeadLatentAttention
from vllm.model_executor.layers.mla import MLAModules, MultiHeadLatentAttention
from vllm.model_executor.layers.quantization import QuantizationConfig
from vllm.model_executor.layers.rotary_embedding import get_rope
from vllm.model_executor.layers.shared_fused_moe import SharedFusedMoE