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 12, 13:30 UTC
Make Transformers Nightly tests soft-fail and enable all tests (#29401)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
Nov 25, 2025 at 12:41 UTC
bf0c75cd4f638b359f52602bb0fd54ef434068cb
1 file changed
+3
-2
.buildkite/test-pipeline.yaml
+3
-2
@@ -903,11 +903,12 @@ steps:
903
- label: Transformers Nightly Models Test
904
working_dir: "/vllm-workspace/"
905
optional: true
906
+ soft_fail: true
907
commands:
908
- pip install --upgrade git+https://github.com/huggingface/transformers
908
- - pytest -v -s tests/models/test_initialization.py -k 'not (Ultravox or Phi4Multimodal or MiniCPMO or Lfm2Moe or RobertaForSequenceClassification or Ovis2_5 or DeepseekOCR or KimiVL)'
909
+ - pytest -v -s tests/models/test_initialization.py
910
- pytest -v -s tests/models/test_transformers.py
910
- # - pytest -v -s tests/models/multimodal/processing/
911
+ - pytest -v -s tests/models/multimodal/processing/
912
- pytest -v -s tests/models/multimodal/test_mapping.py
913
- python3 examples/offline_inference/basic/chat.py
914
- python3 examples/offline_inference/vision_language.py --model-type qwen2_5_vl