obscura/vllm
2
0
Fork 0
mirror of https://github.com/vllm-project/vllm.git synced 2026-08-24 22:50:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

[XPU] limit max-num-seqs in test_lmeval.py for XPU (#47682)

Browse Source 
Signed-off-by: mauyuyuace <[email protected]>
This commit is contained in:
Qiming Zhang
2026-07-06 05:34:16 +00:00
committed by GitHub
parent 69715823df
commit 394edc8108
1 changed files with 3 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/entrypoints/openai/correctness/test_lmeval.py
+3 -2
View File
@@ -71,8 +71,9 @@ def test_lm_eval_accuracy_v1_engine():
more_args = []
# Limit compilation time for V1
if current_platform.is_tpu():
# Limit compilation time for V1 on TPU
# Avoid OOM on XPU
if current_platform.is_tpu() or current_platform.is_xpu():
more_args = ["--max-num-seqs", "64"]
run_test(more_args)