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, 11:00 UTC
Enable standard language model for torhc nightly (#18004)
Signed-off-by: Yang Wang <elainewy@meta.com>
Yang Wang committed
May 12, 2025 at 14:00 UTC
2b0db9b0e2378dbb7c44dd17e4066b1f42d42b70
1 file changed
+4
.buildkite/test-pipeline.yaml
+4
@@ -472,12 +472,14 @@ steps:
472
473
- label: Language Models Test (Standard)
474
mirror_hardwares: [amdexperimental]
475
+ torch_nightly: true
476
source_file_dependencies:
477
- vllm/
478
- tests/models/language
479
commands:
480
# Install causal-conv1d for plamo2 models here, as it is not compatible with pip-compile.
481
- pip install 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.0.post8'
482
+ - pip freeze | grep -E 'torch'
483
- pytest -v -s models/language -m core_model
484
485
- label: Language Models Test (Extended)
@@ -493,11 +495,13 @@ steps:
495
496
- label: Multi-Modal Models Test (Standard)
497
mirror_hardwares: [amdexperimental]
498
+ torch_nightly: true
499
source_file_dependencies:
500
- vllm/
501
- tests/models/multimodal
502
commands:
503
- pip install git+https://github.com/TIGER-AI-Lab/Mantis.git
504
+ - pip freeze | grep -E 'torch'
505
- pytest -v -s models/multimodal/processing
506
- pytest -v -s --ignore models/multimodal/generation/test_whisper.py models/multimodal -m core_model
507
- cd .. && pytest -v -s tests/models/multimodal/generation/test_whisper.py -m core_model # Otherwise, mp_method="spawn" doesn't work