biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bug] Fix attention_backend arg string parsing (#30534)

Browse Source 
Signed-off-by: mgoin <mgoin64@gmail.com>
This commit is contained in:
Michael Goin
2025-12-12 10:40:50 -05:00
committed by GitHub
parent dc13c99eed
commit 09ad3b76b3
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/engine/arg_utils.py
View File

@@ -1649,6 +1649,12 @@ class EngineArgs:
"attention_backend and attention_config.backend " "attention_backend and attention_config.backend "
"are mutually exclusive" "are mutually exclusive"
) )
# Convert string to enum if needed (CLI parsing returns a string)
if isinstance(self.attention_backend, str):
attention_config.backend = AttentionBackendEnum[
self.attention_backend.upper()
]
else:
attention_config.backend = self.attention_backend attention_config.backend = self.attention_backend
load_config = self.create_load_config() load_config = self.create_load_config()