Sign inSign up

vllm/vllm-openai:cu129-nightly-81198e97ba7eee2a22540caaa756b7fdddcb4d93

Multi-platform
Index digest

sha256:56209c8a2a42323b0757b941332ce102c97e5e23a9786b163ce226a573be76cf

OS/ARCH

Compressed size

11.19 GB

Last pushed

18 days by vllmbot

Type

Image

Manifest digest

sha256:3286d41b8a0ec7292f57567fdd6ffb9f6041da01d9c1f3c501994440e4eda4bd

Image Layers

Image Layers

1ARG RELEASE0 B
2ARG LAUNCHPAD_BUILD_ARCH0 B
3LABEL org.opencontainers.image.ref.name=ubuntu0 B
4LABEL org.opencontainers.image.version=24.040 B
5ADD file ... in / 28.55 MB
6CMD ["/bin/bash"]0 B
7ENV NVARCH=sbsa0 B
8ENV NVIDIA_REQUIRE_CUDA=cuda>=12.9 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<5360 B
9ENV NV_CUDA_CUDART_VERSION=12.9.79-10 B
10ARG TARGETARCH0 B
11LABEL maintainer=NVIDIA CORPORATION <[email protected]>0 B
12RUN |1 TARGETARCH=arm64 /bin/sh -c4.32 MB
13ENV CUDA_VERSION=12.9.10 B
14RUN |1 TARGETARCH=arm64 /bin/sh -c93.06 MB
15RUN |1 TARGETARCH=arm64 /bin/sh -c182 B
16ENV PATH=/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
17ENV LD_LIBRARY_PATH=/usr/local/cuda/lib640 B
18COPY NGC-DL-CONTAINER-LICENSE / # buildkit6.72 KB
19ENV NVIDIA_VISIBLE_DEVICES=all0 B
20ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility0 B
21ARG INSTALL_RUBIN_PRERELEASE0 B
22ARG TARGETPLATFORM0 B
23ARG max_jobs=20 B
24ARG CUDA_VERSION0 B
25ARG PYTHON_VERSION0 B
26ARG NCCL_VERSION0 B
27ARG DEADSNAKES_MIRROR_URL0 B
28ARG DEADSNAKES_GPGKEY_URL0 B
29ARG GET_PIP_URL0 B
30ENV DEBIAN_FRONTEND=noninteractive0 B
31WORKDIR /vllm-workspace104 B
32RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16229 B
33RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16216.18 MB
34RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=161.56 GB
35ENV CUDA_HOME=/usr/local/cuda0 B
36ENV PATH=/usr/local/cuda/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
37ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib640 B
38COPY tools/install_snapshot_runtime.sh /usr/local/bin/install_snapshot_runtime.sh # buildkit1.62 KB
39RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=1632 B
40ENV CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu0 B
41RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=1638.67 MB
42ENV UV_HTTP_TIMEOUT=5000 B
43ENV UV_INDEX_STRATEGY=unsafe-best-match0 B
44ENV UV_LINK_MODE=copy0 B
45ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python0 B
46ENV UV_CACHE_DIR=/opt/uv/cache0 B
47ENV UV_COMPILE_BYTECODE=10 B
48RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16151 B
49RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16164 B
50ENV UV_OVERRIDE=/etc/uv-overrides.txt0 B
51RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=163.62 KB
52COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit2.03 KB
53RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=162.03 KB
54ENV VLLM_ENABLE_CUDA_COMPATIBILITY=00 B
55ARG PYTORCH_CUDA_INDEX_BASE_URL0 B
56COPY requirements/common.txt /tmp/common.txt # buildkit1.67 KB
57COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit867 B
58RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=165.83 GB
59ARG FLASHINFER_VERSION=0.7.0.post10 B
60RUN |11 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=161.14 GB
61ARG GDRCOPY_CUDA_VERSION=12.80 B
62ARG GDRCOPY_OS_VERSION=Ubuntu24_040 B
63ARG TARGETPLATFORM0 B
64COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit897 B
65RUN |14 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16517.71 KB
66ARG TIMM_VERSION=>=1.0.170 B
67ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.70 B
68RUN |16 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16303.76 MB
69ARG PIP_INDEX_URL UV_INDEX_URL0 B
70ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL0 B
71ARG PYTORCH_CUDA_INDEX_BASE_URL0 B
72ARG PIP_KEYRING_PROVIDER UV_KEYRING_PROVIDER0 B
73ARG PYTORCH_NIGHTLY0 B
74COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit196 B
75COPY /workspace/dist/wheel.sha256 /tmp/vllm-wheel.sha256 # buildkit243 B
76RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=749.63 MB
77RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=199 B
78RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=32 B
79ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib640 B
80COPY /tmp/ep_kernels_workspace/dist/wheels.sha256 /tmp/ep-kernels-wheels.sha256 # buildkit254 B
81RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=18.09 MB
82ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib640 B
83COPY examples examples # buildkit308.57 KB
84COPY benchmarks benchmarks # buildkit294.09 KB
85COPY ./vllm/collect_env.py . # buildkit9.61 KB
86ARG TARGETPLATFORM0 B
87ARG INSTALL_KV_CONNECTORS=false0 B
88ARG INSTALL_RUBIN_PRERELEASE0 B
89ARG TRITON_INSTALL_FROM_SOURCE_REPO0 B
90ARG CUDA_VERSION0 B
91ARG VLLM_BUILD_COMMIT0 B
92ARG VLLM_BUILD_PIPELINE0 B
93ARG VLLM_BUILD_URL0 B
94ARG VLLM_IMAGE_TAG0 B
95ARG PIP_INDEX_URL UV_INDEX_URL0 B
96ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL0 B
97ENV UV_HTTP_TIMEOUT=5000 B
98ARG torch_cuda_arch_list=7.5 8.0 8.6 8.90 B
99ENV TORCH_CUDA_ARCH_LIST=8.0 8.7 8.9 9.00 B
100RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false1.25 GB
101RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false92 B
102COPY /tmp/triton_from_source_workspace/dist/wheel.sha256 /tmp/triton-source-wheel.sha256 # buildkit212 B
103COPY tools/install_triton_from_source.sh /tmp/install_triton_from_source.sh # buildkit824 B
104RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false183 B
105ARG BUILD_NIXL0 B
106RUN |15 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false92 B
107ENV VLLM_USAGE_SOURCE=production-docker-image0 B
108ENV VLLM_BUILD_COMMIT=81198e97ba7eee2a22540caaa756b7fdddcb4d93 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/7974 VLLM_IMAGE_TAG=vllm/vllm-openai:cu129-nightly-81198e97ba7eee2a22540caaa756b7fdddcb4d930 B
109LABEL org.opencontainers.image.source=https://github.com/vllm-project/vllm org.opencontainers.image.revision=81198e97ba7eee2a22540caaa756b7fdddcb4d93 org.opencontainers.image.version=vllm/vllm-openai:cu129-nightly-81198e97ba7eee2a22540caaa756b7fdddcb4d93 org.opencontainers.image.url=https://buildkite.com/vllm/release-v2/builds/79740 B
110ENTRYPOINT ["vllm" "serve"]0 B

Command

 ARG RELEASE