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

[platforms] restore xpu check for parallel config (#10479)

Browse Source 
Signed-off-by: youkaichao <youkaichao@gmail.com>
This commit is contained in:
youkaichao 2024-11-20 09:13:28 -08:00 committed by GitHub
parent 63f1fde277
commit 772a66732d
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
vllm/platforms/xpu.py
View File

@ -55,3 +55,13 @@ class XPUPlatform(Platform):
"CUDA graph is not supported on XPU, fallback to the eager " "CUDA graph is not supported on XPU, fallback to the eager "
"mode.") "mode.")
model_config.enforce_eager = True model_config.enforce_eager = True
# check and update parallel config
parallel_config = vllm_config.parallel_config
if (parallel_config.distributed_executor_backend is not None
and parallel_config.distributed_executor_backend != "ray"):
logger.warning(
"%s is not supported on XPU, fallback to ray distributed"
" executor backend.",
parallel_config.distributed_executor_backend)
parallel_config.distributed_executor_backend = "ray"