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

[V1] TPU CI - Add basic perf regression test (#15414)

Browse Source 
Signed-off-by: Alexander Matveev <amatveev@redhat.com>
This commit is contained in:
Alexander Matveev
2025-03-31 13:25:20 -04:00
committed by GitHub
parent 2de4118243
commit 9a2160fa55
5 changed files with 192 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/entrypoints/llm/test_accuracy.py
View File

@@ -58,7 +58,7 @@ def test_lm_eval_accuracy_v1_engine(monkeypatch: pytest.MonkeyPatch):
more_args = None
if current_platform.is_tpu():
# Limit compilation time for TPU V1
more_args = "max_num_seqs=64"
more_args = "max_model_len=2048,max_num_seqs=64"
# Add TP test (if provided)
if TPU_TP_TEST_STR: