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

fix qwenvl pd smoke test error (#4597)

Browse Source 
### What this PR does / why we need it?
Fix A3 QwenVL PD smoke test error
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?
curl a requet to proxy port，it responses correctly

- vLLM version: v0.11.2

Signed-off-by: 李少鹏 <lishaopeng21@huawei.com>
This commit is contained in:
shaopeng-666
2025-12-01 22:24:59 +08:00
committed by GitHub
parent 143e1f46d0
commit 8e7f5cff6d
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
vllm_ascend/distributed/llmdatadist_c_mgr_connector.py
View File

@@ -107,6 +107,7 @@ class LLMDataDistCMgrConnector(KVConnectorBase_V1):
kv_cache_config: Optional[KVCacheConfig] = None): kv_cache_config: Optional[KVCacheConfig] = None):
assert vllm_config.kv_transfer_config is not None assert vllm_config.kv_transfer_config is not None
self.engine_id = vllm_config.kv_transfer_config.engine_id self.engine_id = vllm_config.kv_transfer_config.engine_id
self._connector_metadata = LLMDataDistCMgrConnectorMetadata()
if role == KVConnectorRole.SCHEDULER: if role == KVConnectorRole.SCHEDULER:
self.connector_scheduler: Optional[ self.connector_scheduler: Optional[
LLMDataDistCMgrConnectorScheduler] = LLMDataDistCMgrConnectorScheduler( LLMDataDistCMgrConnectorScheduler] = LLMDataDistCMgrConnectorScheduler(