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 10, 12:45 UTC
[docker] A follow-up patch to fix #30913: `[docker] install cuda13 version of lmcache and nixl` (#31775)
Signed-off-by: Shang Wang <shangw@nvidia.com>
Shang Wang committed
Jan 8, 2026 at 02:47 UTC
33156f56e0b1a4e872048669ae625fe3067f9c24
1 file changed
+2
docker/Dockerfile
+2
@@ -627,6 +627,8 @@ ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL
627
ENV UV_HTTP_TIMEOUT=500
628
629
# install kv_connectors if requested
630
+ARG torch_cuda_arch_list='7.0 7.5 8.0 8.9 9.0 10.0 12.0'
631
+ENV TORCH_CUDA_ARCH_LIST=${torch_cuda_arch_list}
632
RUN --mount=type=cache,target=/root/.cache/uv \
633
--mount=type=bind,source=requirements/kv_connectors.txt,target=/tmp/kv_connectors.txt,ro \
634
CUDA_MAJOR="${CUDA_VERSION%%.*}"; \