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

[Misc] Remove redundant imported envs, using envs_ascend instead (#2193)

Browse Source 
### What this PR does / why we need it?
Remove redundant imported `envs`, using `envs_ascend` instead.

```python
import vllm.envs as envs_vllm
import vllm_ascend.envs as envs_ascend
```

- vLLM version: v0.10.0
- vLLM main:
71683ca6f6

---------

Signed-off-by: shen-shanshan <467638484@qq.com>
This commit is contained in:
Shanshan Shen
2025-08-14 09:33:39 +08:00
committed by GitHub
parent 55d0790597
commit 103654ccd6
14 changed files with 46 additions and 46 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm_ascend/attention/mla_v1.py
View File

@@ -14,7 +14,7 @@ from vllm.model_executor.layers.linear import (LinearBase,
UnquantizedLinearMethod)
from vllm.utils import cdiv, round_down
from vllm_ascend import envs
import vllm_ascend.envs as envs_ascend
from vllm_ascend.ascend_config import get_ascend_config
from vllm_ascend.attention.attention_v1 import AscendAttentionState
from vllm_ascend.multistream.base import MSAttentionMetadataSplitConfig
@@ -1054,7 +1054,7 @@ class AscendMLAImpl(MLAAttentionImpl):
# be removed after the torch_npu contains `torch_npu.atb.npu_multi_head_latent_attention` become
# public available
assert len(kv_c_and_k_pe_cache) > 1
if envs.VLLM_ASCEND_MLA_PA:
if envs_ascend.VLLM_ASCEND_MLA_PA:
attn_output = torch_npu.atb.npu_multi_head_latent_attention(
q_nope, q_pe, kv_c_and_k_pe_cache[0],
kv_c_and_k_pe_cache[1], attn_metadata.decode.block_table,