vllm/.buildkite/test_areas/models_basic.yaml
2025-12-08 17:25:43 -09:00

63 lines
2.3 KiB
YAML

group: Models - Basic
depends_on:
- image-build
steps:
- label: Basic Models Tests (Initialization)
timeout_in_minutes: 45
mirror_hardwares: [amdexperimental]
torch_nightly: true
source_file_dependencies:
- vllm/
- tests/models/test_initialization.py
commands:
# Run a subset of model initialization tests
- pytest -v -s models/test_initialization.py::test_can_initialize_small_subset
- label: Basic Models Tests (Extra Initialization) %N
timeout_in_minutes: 45
mirror_hardwares: [amdexperimental]
torch_nightly: true
source_file_dependencies:
- vllm/model_executor/models/
- tests/models/test_initialization.py
commands:
# Only when vLLM model source is modified - test initialization of a large
# subset of supported models (the complement of the small subset in the above
# test.) Also run if model initialization test file is modified
- pytest -v -s models/test_initialization.py -k 'not test_can_initialize_small_subset' --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --shard-id=$$BUILDKITE_PARALLEL_JOB
parallelism: 2
- label: Basic Models Tests (Other)
timeout_in_minutes: 45
source_file_dependencies:
- vllm/
- tests/models/test_transformers.py
- tests/models/test_registry.py
commands:
- pytest -v -s models/test_transformers.py models/test_registry.py
- label: Basic Models Test (Other CPU) # 5min
timeout_in_minutes: 10
source_file_dependencies:
- vllm/
- tests/models/test_utils.py
- tests/models/test_vision.py
no_gpu: true
commands:
- pytest -v -s models/test_utils.py models/test_vision.py
- label: Transformers Nightly Models
working_dir: "/vllm-workspace/"
optional: true
soft_fail: true
commands:
- pip install --upgrade git+https://github.com/huggingface/transformers
- pytest -v -s tests/models/test_initialization.py
- pytest -v -s tests/models/test_transformers.py
- pytest -v -s tests/models/multimodal/processing/
- pytest -v -s tests/models/multimodal/test_mapping.py
- python3 examples/offline_inference/basic/chat.py
- python3 examples/offline_inference/vision_language.py --model-type qwen2_5_vl
# Whisper needs spawn method to avoid deadlock
- VLLM_WORKER_MULTIPROC_METHOD=spawn python3 examples/offline_inference/audio_language.py --model-type whisper