[V1] V1 Enablement Oracle (#13726)

Signed-off-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
Co-authored-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
This commit is contained in:
Robert Shaw
2025-03-15 01:02:20 -04:00
committed by GitHub
parent 8c0d15d5c5
commit d4d93db2c5
96 changed files with 1537 additions and 512 deletions

View File

@@ -111,6 +111,26 @@ VIDEO_ASSETS = _VideoAssets()
"""Singleton instance of :class:`_VideoAssets`."""
@pytest.fixture(scope="function", autouse=True)
def cleanup_VLLM_USE_V1(monkeypatch):
"""
The V1 oracle sets "VLLM_USE_V1" during loading. This means
that each invocation of a test change the env variable.
If we touch "VLLM_USE_V1" with monkeypatch, then any changes
made during the test run by vLLM will be cleaned up.
This fixture is used by every test.
"""
# If VLLM_USE_V1 is not set, set then delete. This will
# cause monkeypatch to clean up VLLM_USE_V1 upon exit
# if VLLM modifies the value of envs.VLLM_USE_V1.
if "VLLM_USE_V1" not in os.environ:
monkeypatch.setenv("VLLM_USE_V1", "")
monkeypatch.delenv("VLLM_USE_V1")
@pytest.fixture(params=[True, False])
def run_with_both_engines(request, monkeypatch):
# Automatically runs tests twice, once with V1 and once without