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

[long_seq] remove long_seq env (#4660)

Browse Source 
### What this PR does / why we need it?
remove env VLLM_ASCEND_ENABLE_CONTEXT_PARALLEL 

- vLLM version: v0.12.0

---------

Signed-off-by: LookAround <lixushi@huawei.com>
Signed-off-by: ZhangMingWei716 <2894054457@qq.com>
Co-authored-by: ZhangMingWei716 <2894054457@qq.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>
This commit is contained in:
LookAround0301
2025-12-05 10:31:49 +08:00
committed by GitHub
parent ea54388e19
commit b32ef53b3b
16 changed files with 230 additions and 176 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
vllm_ascend/ops/fused_moe/prepare_finalize.py
View File

@@ -24,16 +24,13 @@ import torch.nn as nn
import torch_npu
from vllm.distributed import tensor_model_parallel_all_reduce
from vllm.distributed.parallel_state import (
get_dp_group, get_tensor_model_parallel_rank,
get_dp_group, get_pcp_group, get_tensor_model_parallel_rank,
get_tensor_model_parallel_world_size)
from vllm.forward_context import get_forward_context
from vllm.model_executor.layers.fused_moe import FusedMoEConfig
from vllm_ascend.utils import enable_sp, prefill_context_parallel_enable
if prefill_context_parallel_enable():
from vllm.distributed import get_pcp_group
class QuantType(Enum):
NONE = 0