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 2, 17:01 UTC
Remove Transformers forward/backward compatibility tests (#43785)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
May 27, 2026 at 20:46 UTC
2616f67faaa735a3e0d9c17968fa91f242d36c56
1 file changed
-34
.buildkite/test_areas/models_basic.yaml
-34
@@ -58,37 +58,3 @@ steps:
58
device: cpu-small
59
commands:
60
- pytest -v -s models/test_utils.py models/test_vision.py
61
-
62
-- label: Transformers Nightly Models
63
- device: h200_35gb
64
- key: transformers-nightly-models
65
- working_dir: "/vllm-workspace/"
66
- optional: true
67
- soft_fail: true
68
- commands:
69
- - pip install --upgrade git+https://github.com/huggingface/transformers
70
- - pytest -v -s tests/models/test_initialization.py
71
- - pytest -v -s tests/models/test_transformers.py
72
- - pytest -v -s tests/models/multimodal/processing/
73
- - pytest -v -s tests/models/multimodal/test_mapping.py
74
- - python3 examples/basic/offline_inference/chat.py
75
- - python3 examples/generate/multimodal/vision_language_offline.py --model-type qwen2_5_vl
76
- # Whisper needs spawn method to avoid deadlock
77
- - VLLM_WORKER_MULTIPROC_METHOD=spawn python3 examples/generate/multimodal/audio_language_offline.py --model-type whisper
78
-
79
-- label: Transformers Backward Compatibility Models Test
80
- device: h200_35gb
81
- key: transformers-backward-compatibility-models-test
82
- working_dir: "/vllm-workspace/"
83
- optional: true
84
- soft_fail: true
85
- commands:
86
- - pip install transformers==4.57.5
87
- - pytest -v -s tests/models/test_initialization.py
88
- - pytest -v -s tests/models/test_transformers.py
89
- - pytest -v -s tests/models/multimodal/processing/
90
- - pytest -v -s tests/models/multimodal/test_mapping.py
91
- - python3 examples/basic/offline_inference/chat.py
92
- - python3 examples/generate/multimodal/vision_language_offline.py --model-type qwen2_5_vl
93
- # Whisper needs spawn method to avoid deadlock
94
- - VLLM_WORKER_MULTIPROC_METHOD=spawn python3 examples/generate/multimodal/audio_language_offline.py --model-type whisper