A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 03:30 UTC
[Misc] Remove the unused LoRA test code (#20494)
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
Jee Jee Li committed
Jul 5, 2025 at 13:48 UTC
906e05d8405f682797b5c934ec84d3c272a5fafd
2 files changed
+1
-18
tests/lora/conftest.py
-17
@@ -249,23 +249,6 @@ def llama_2_7b_model_extra_embeddings(llama_2_7b_engine_extra_embeddings):
249
model_runner.model)
250
251
252
-@pytest.fixture(params=[True, False])
253
-def run_with_both_engines_lora(request, monkeypatch):
254
- # Automatically runs tests twice, once with V1 and once without
255
- use_v1 = request.param
256
- # Tests decorated with `@skip_v1` are only run without v1
257
- skip_v1 = request.node.get_closest_marker("skip_v1")
258
-
259
- if use_v1:
260
- if skip_v1:
261
- pytest.skip("Skipping test on vllm V1")
262
- monkeypatch.setenv('VLLM_USE_V1', '1')
263
- else:
264
- monkeypatch.setenv('VLLM_USE_V1', '0')
265
-
266
- yield
267
-
268
-
252
@pytest.fixture
253
def reset_default_device():
254
"""
vllm/multimodal/utils.py
+1
-1
@@ -489,4 +489,4 @@ def fetch_video(
489
"video": video_io_kwargs
490
}
491
media_connector = MediaConnector(media_io_kwargs=media_io_kwargs)
492
- return media_connector.fetch_video(video_url)
492
+ return media_connector.fetch_video(video_url)
\ No newline at end of file