Sign inSign up

vllm/vllm-openai:minimax-m31-flash

Multi-platform
Index digest

sha256:86f582dae25ea92802353727164ae4eb03b37b18baf9119b805ffaec2a7ef07c

OS/ARCH

Compressed size

10.73 GB

Last pushed

14 days by vllmbot

Type

Image

Manifest digest

sha256:5d76347628973fb80519a919a2f2740b6645b62cba3fe2bd87b25f111bd0e948

Image Layers

Image Layers

1ARG RELEASE0 B
2ARG LAUNCHPAD_BUILD_ARCH0 B
3LABEL org.opencontainers.image.version=24.040 B
4ADD file ... in / 28.62 MB
5CMD ["/bin/bash"]0 B
6ENV NVARCH=sbsa0 B
7ENV NVIDIA_REQUIRE_CUDA=cuda>=13.4 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<5360 B
8ENV NV_CUDA_CUDART_VERSION=13.4.49-10 B
9ARG TARGETARCH0 B
10LABEL maintainer=NVIDIA CORPORATION <[email protected]>0 B
11ARG IMAGE_UPGRADE_ID=00 B
12RUN |2 TARGETARCH=arm64 IMAGE_UPGRADE_ID=20260914 /bin/sh9.9 MB
13ENV CUDA_VERSION=13.4.10 B
14RUN |2 TARGETARCH=arm64 IMAGE_UPGRADE_ID=20260914 /bin/sh398.13 KB
15RUN |2 TARGETARCH=arm64 IMAGE_UPGRADE_ID=20260914 /bin/sh153.01 MB
16RUN |2 TARGETARCH=arm64 IMAGE_UPGRADE_ID=20260914 /bin/sh152.92 MB
17RUN |2 TARGETARCH=arm64 IMAGE_UPGRADE_ID=20260914 /bin/sh187 B
18ENV PATH=/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
19ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
20COPY NGC-DL-CONTAINER-LICENSE / # buildkit6.72 KB
21ENV NVIDIA_VISIBLE_DEVICES=all0 B
22ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility0 B
23ARG INSTALL_RUBIN_PRERELEASE=false0 B
24ARG TARGETPLATFORM=linux/arm640 B
25ARG max_jobs=640 B
26ARG CUDA_VERSION=13.4.10 B
27ARG PYTHON_VERSION=3.120 B
28ARG NCCL_VERSION=2.30.70 B
29ARG DEADSNAKES_MIRROR_URL0 B
30ARG DEADSNAKES_GPGKEY_URL0 B
31ARG GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py0 B
32ENV DEBIAN_FRONTEND=noninteractive0 B
33WORKDIR /vllm-workspace104 B
34RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64230 B
35RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64208.01 MB
36RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=641.54 GB
37ENV CUDA_HOME=/usr/local/cuda0 B
38ENV 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
39ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
40COPY tools/install_snapshot_runtime.sh /usr/local/bin/install_snapshot_runtime.sh # buildkit1.62 KB
41RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
42ENV CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu0 B
43RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432.28 MB
44ENV UV_HTTP_TIMEOUT=5000 B
45ENV UV_INDEX_STRATEGY=unsafe-best-match0 B
46ENV UV_LINK_MODE=copy0 B
47ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python0 B
48ENV UV_CACHE_DIR=/opt/uv/cache0 B
49ENV UV_COMPILE_BYTECODE=10 B
50RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64150 B
51RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64164 B
52ENV UV_OVERRIDE=/etc/uv-overrides.txt0 B
53RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=643.39 KB
54COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit2.03 KB
55RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
56ENV VLLM_ENABLE_CUDA_COMPATIBILITY=00 B
57ARG PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl0 B
58COPY requirements/common.txt /tmp/common.txt # buildkit1.67 KB
59COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit719 B
60RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=643.89 GB
61RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=642.93 GB
62ARG GDRCOPY_CUDA_VERSION=12.80 B
63ARG GDRCOPY_OS_VERSION=Ubuntu24_040 B
64ARG TARGETPLATFORM=linux/arm640 B
65COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit897 B
66RUN |12 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64316.73 KB
67ARG TIMM_VERSION=>=1.0.170 B
68ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.70 B
69RUN |14 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64274.7 MB
70ARG PIP_INDEX_URL UV_INDEX_URL=0 B
71ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL=0 B
72ARG PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl0 B
73ARG PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled0 B
74ARG PYTORCH_NIGHTLY=10 B
75COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit201 B
76COPY /workspace/dist/wheel.sha256 /tmp/vllm-wheel.sha256 # buildkit246 B
77RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=641018.21 MB
78RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64200 B
79RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
80ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
81COPY /tmp/ep_kernels_workspace/dist/wheels.sha256 /tmp/ep-kernels-wheels.sha256 # buildkit252 B
82RUN |21 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6413.79 MB
83ENV 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
84COPY examples examples # buildkit309.71 KB
85COPY benchmarks benchmarks # buildkit302.15 KB
86COPY ./vllm/collect_env.py . # buildkit9.61 KB
87ARG TARGETPLATFORM=linux/arm640 B
88ARG INSTALL_KV_CONNECTORS=true0 B
89ARG INSTALL_RUBIN_PRERELEASE=false0 B
90ARG TRITON_INSTALL_FROM_SOURCE_REPO=0 B
91ARG CUDA_VERSION=13.4.10 B
92ARG VLLM_BUILD_COMMIT0 B
93ARG VLLM_BUILD_PIPELINE0 B
94ARG VLLM_BUILD_URL0 B
95ARG VLLM_IMAGE_TAG0 B
96ARG PIP_INDEX_URL UV_INDEX_URL=0 B
97ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL=0 B
98ENV UV_HTTP_TIMEOUT=5000 B
99ARG torch_cuda_arch_list=7.5 8.0 8.6 8.90 B
100ENV TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 8.90 B
101RUN |28 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64529.72 MB
102RUN |28 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
103COPY /tmp/triton_from_source_workspace/dist/wheel.sha256 /tmp/triton-source-wheel.sha256 # buildkit212 B
104COPY tools/install_triton_from_source.sh /tmp/install_triton_from_source.sh # buildkit825 B
105RUN |28 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=64174 B
106ARG BUILD_NIXL=false0 B
107RUN |29 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=6432 B
108ENV VLLM_USAGE_SOURCE=production-docker-image0 B
109ENV VLLM_BUILD_COMMIT=unknown VLLM_BUILD_PIPELINE=local VLLM_BUILD_URL= VLLM_IMAGE_TAG=local/vllm-openai:dev0 B
110LABEL 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
111ENTRYPOINT ["vllm" "serve"]0 B
112ENV TRITON_PTXAS_PATH=/usr/local/cuda/bin/ptxas0 B
113RUN /bin/sh -c pip uninstall806 B

Command

 ARG RELEASE