Sign inSign up

vllm/vllm-openai:flint

Multi-platform
Index digest

sha256:be337371b88c770a25a7c029ecc1096dde132889853609b28093ddabc002b0bd

OS/ARCH

Compressed size

10.01 GB

Last pushed

9 days by vllmbot

Type

Image

Manifest digest

sha256:5641d14b70c991f8b78c3a5baf0281cb73dbd2ea69932dd6a328820f420ec4bc

Image Layers

Image Layers

1ARG RELEASE0 B
2ARG LAUNCHPAD_BUILD_ARCH0 B
3LABEL org.opencontainers.image.version=24.040 B
4ADD file ... in / 28.56 MB
5CMD ["/bin/bash"]0 B
6ENV NVARCH=sbsa0 B
7ENV NVIDIA_REQUIRE_CUDA=cuda>=13.0 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<5360 B
8ENV NV_CUDA_CUDART_VERSION=13.0.96-10 B
9ARG TARGETARCH0 B
10LABEL maintainer=NVIDIA CORPORATION <[email protected]>0 B
11RUN |1 TARGETARCH=arm64 /bin/sh -c6.44 MB
12ENV CUDA_VERSION=13.0.30 B
13RUN |1 TARGETARCH=arm64 /bin/sh -c367.96 KB
14RUN |1 TARGETARCH=arm64 /bin/sh -c94.84 MB
15RUN |1 TARGETARCH=arm64 /bin/sh -c32 B
16RUN |1 TARGETARCH=arm64 /bin/sh -c187 B
17ENV PATH=/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
18ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
19COPY NGC-DL-CONTAINER-LICENSE / # buildkit6.72 KB
20ENV NVIDIA_VISIBLE_DEVICES=all0 B
21ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility0 B
22ARG INSTALL_RUBIN_PRERELEASE=false0 B
23ARG TARGETPLATFORM=linux/arm640 B
24ARG max_jobs=640 B
25ARG CUDA_VERSION=13.0.30 B
26ARG PYTHON_VERSION=3.120 B
27ARG NCCL_VERSION=2.30.70 B
28ARG DEADSNAKES_MIRROR_URL0 B
29ARG DEADSNAKES_GPGKEY_URL0 B
30ARG GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py0 B
31ENV DEBIAN_FRONTEND=noninteractive0 B
32WORKDIR /vllm-workspace104 B
33RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64230 B
34RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64215.17 MB
35RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=641.49 GB
36ENV CUDA_HOME=/usr/local/cuda0 B
37ENV PATH=/usr/local/cuda/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
38ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
39COPY tools/install_snapshot_runtime.sh /usr/local/bin/install_snapshot_runtime.sh # buildkit1.62 KB
40RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
41ENV CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu0 B
42RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432.28 MB
43ENV UV_HTTP_TIMEOUT=5000 B
44ENV UV_INDEX_STRATEGY=unsafe-best-match0 B
45ENV UV_LINK_MODE=copy0 B
46ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python0 B
47ENV UV_CACHE_DIR=/opt/uv/cache0 B
48ENV UV_COMPILE_BYTECODE=10 B
49RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64150 B
50RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64164 B
51ENV UV_OVERRIDE=/etc/uv-overrides.txt0 B
52RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=643.4 KB
53COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit2.04 KB
54RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
55ENV VLLM_ENABLE_CUDA_COMPATIBILITY=00 B
56ARG PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl0 B
57COPY requirements/common.txt /tmp/common.txt # buildkit1.67 KB
58COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit718 B
59RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=643.89 GB
60RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=642.93 GB
61ARG GDRCOPY_CUDA_VERSION=12.80 B
62ARG GDRCOPY_OS_VERSION=Ubuntu24_040 B
63ARG TARGETPLATFORM=linux/arm640 B
64COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit897 B
65RUN |12 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64316.69 KB
66ARG TIMM_VERSION=>=1.0.170 B
67ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.70 B
68RUN |14 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64274.7 MB
69ARG PIP_INDEX_URL UV_INDEX_URL=0 B
70ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL=0 B
71ARG PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl0 B
72ARG PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled0 B
73ARG PYTORCH_NIGHTLY0 B
74COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit197 B
75COPY /workspace/dist/wheel.sha256 /tmp/vllm-wheel.sha256 # buildkit241 B
76RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64550.09 MB
77RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64200 B
78RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
79ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
80COPY /tmp/ep_kernels_workspace/dist/wheels.sha256 /tmp/ep-kernels-wheels.sha256 # buildkit253 B
81RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6413.81 MB
82ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
83COPY examples examples # buildkit309.71 KB
84COPY benchmarks benchmarks # buildkit300.29 KB
85COPY ./vllm/collect_env.py . # buildkit9.61 KB
86ARG TARGETPLATFORM=linux/arm640 B
87ARG INSTALL_KV_CONNECTORS=true0 B
88ARG INSTALL_RUBIN_PRERELEASE=false0 B
89ARG TRITON_INSTALL_FROM_SOURCE_REPO=0 B
90ARG CUDA_VERSION=13.0.30 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_URL=0 B
96ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL=0 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=7.5 8.0 8.6 8.90 B
100RUN |28 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64529.73 MB
101RUN |28 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 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 # buildkit825 B
104RUN |28 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64173 B
105ARG BUILD_NIXL=false0 B
106RUN |29 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
107ENV VLLM_USAGE_SOURCE=production-docker-image0 B
108ENV VLLM_BUILD_COMMIT=unknown VLLM_BUILD_PIPELINE=local VLLM_BUILD_URL= VLLM_IMAGE_TAG=local/vllm-openai:dev0 B
109LABEL org.opencontainers.image.source=https://github.com/vllm-project/vllm org.opencontainers.image.revision=unknown org.opencontainers.image.version=local/vllm-openai:dev org.opencontainers.image.url=0 B
110ENTRYPOINT ["vllm" "serve"]0 B

Command

 ARG RELEASE