[Feat] Support Pipeline Resolution Based on HF Config & Make Pipeline Registry Eager (#3760)

Signed-off-by: Alex Brooks <albrooks@redhat.com>
This commit is contained in:
Alex Brooks authored and GitHub committed 2026-06-24 11:02:29 +08:00
1 parent f6894f4308
commit 465f1464d2
44 files changed
+1038 -1126

No files matched your search

+17 -1
View File
@@ -14,7 +14,7 @@ from tests.helpers.stage_config import get_deploy_config_path, modify_stage_conf
from vllm_omni.platforms import current_omni_platform
models = ["Qwen/Qwen3-Omni-30B-A3B-Instruct"]
thinker_only_models = ["Qwen/Qwen3-Omni-30B-A3B-Captioner"]
# Single CI deploy YAML; rocm/xpu deltas are picked automatically via the
# platforms: section. Only CUDA needs an extra enforce_eager tweak.
@@ -40,6 +40,10 @@ else:
# Create parameter combinations for model and stage config
test_params = [(model, stage_config) for model in models for stage_config in stage_configs]
# we can use the same config for a model that only has thinker (i.e., does not
# enable audio output) because the resolver should figure out that it doesn't
# need the full pipeline based on the HF config.
thinker_test_params = [(model, stage_config) for model in thinker_only_models for stage_config in stage_configs]
def get_question(prompt_type="video"):
@@ -61,3 +65,15 @@ def test_video_to_audio(omni_runner, omni_runner_handler) -> None:
# Test single completion
omni_runner_handler.send_omni_request(request_config)
@pytest.mark.advanced_model
@pytest.mark.omni
@hardware_test(res={"cuda": "H100", "rocm": "MI325"}, num_cards=1)
@pytest.mark.parametrize("omni_runner", thinker_test_params, indirect=True)
def test_thinker_only_model_request(omni_runner, omni_runner_handler) -> None:
"""Test that we can load and run a request through a model that only has the thinker stage."""
request_config = {"prompts": "what color is the sky?", "modalities": ["text"]}
# Test single completion
omni_runner_handler.send_omni_request(request_config)