diff --git a/qa/L0_e2e/test_model.py b/qa/L0_e2e/test_model.py index 4ceb976..d2d9453 100644 --- a/qa/L0_e2e/test_model.py +++ b/qa/L0_e2e/test_model.py @@ -68,6 +68,7 @@ def valid_shm_modes(): # TODO(hcho3): Remove once we fix the flakiness of CUDA shared mem +# See https://github.com/triton-inference-server/server/issues/7688 def shared_mem_parametrize(): params = [None] if "cuda" in valid_shm_modes(): diff --git a/qa/run_tests.sh b/qa/run_tests.sh index 8eeddda..6f5b6ee 100755 --- a/qa/run_tests.sh +++ b/qa/run_tests.sh @@ -102,10 +102,9 @@ trap finally EXIT if [ ! -z $CPU_ONLY ] && [ $CPU_ONLY -eq 1 ] then pytest \ - -v -rxXs \ --repo "${MODEL_REPO}" \ --hypothesis-profile "$TEST_PROFILE" \ "$QA_DIR" else - pytest -v -rxXs --repo "${MODEL_REPO}" "$QA_DIR" --hypothesis-profile "$TEST_PROFILE" + pytest --repo "${MODEL_REPO}" "$QA_DIR" --hypothesis-profile "$TEST_PROFILE" fi