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 14, 05:30 UTC
[CI/Build] Update causal-conv1d installation (#27529)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Oct 26, 2025 at 22:14 UTC
55cba4a05c203509b18e9c792f24d77bbe840f31
2 files changed
+8
-4
.buildkite/test-amd.yaml
+4
-2
@@ -789,8 +789,10 @@ steps:
789
- vllm/
790
- tests/models/language/generation
791
commands:
792
- # Install causal-conv1d for plamo2 models here, as it is not compatible with pip-compile.
793
- - pip install 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.0.post8'
792
+ # Install fast path packages for testing against transformers
793
+ # Note: also needed to run plamo2 model in vLLM
794
+ - uv pip install --system --no-build-isolation 'git+https://github.com/state-spaces/mamba@v2.2.5'
795
+ - uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.2'
796
- pytest -v -s models/language/generation -m '(not core_model) and (not hybrid_model)'
797
798
- label: Language Models Test (PPL)
.buildkite/test-pipeline.yaml
+4
-2
@@ -699,8 +699,10 @@ steps:
699
- vllm/
700
- tests/models/language/generation
701
commands:
702
- # Install causal-conv1d for plamo2 models here, as it is not compatible with pip-compile.
703
- - pip install 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.0.post8'
702
+ # Install fast path packages for testing against transformers
703
+ # Note: also needed to run plamo2 model in vLLM
704
+ - uv pip install --system --no-build-isolation 'git+https://github.com/state-spaces/mamba@v2.2.5'
705
+ - uv pip install --system --no-build-isolation 'git+https://github.com/Dao-AILab/causal-conv1d@v1.5.2'
706
- pytest -v -s models/language/generation -m '(not core_model) and (not hybrid_model)'
707
708
- label: Language Models Test (PPL)