Sign inSign up

vllm/vllm-openai:deepseekv4-flash-vision

Multi-platform
Index digest

sha256:0075fd82e3b6d943b0aa91e35da8dbca63d88516c607745131055e9d81f37ebb

OS/ARCH

Compressed size

9.01 GB

Last pushed

18 days by vllmbot

Type

Image

Manifest digest

sha256:8568b4bbc821903d93a0a9c17dd80382fdc0ba78eaa128e3eb5cb71c3bf06b79

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 / 28.56 MB
6CMD ["/bin/bash"]0 B
7ENV NVARCH=sbsa0 B
8ENV 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
9ENV NV_CUDA_CUDART_VERSION=13.0.88-10 B
10ARG TARGETARCH0 B
11LABEL maintainer=NVIDIA CORPORATION <[email protected]>0 B
12RUN |1 TARGETARCH=arm64 /bin/sh -c4.32 MB
13ENV CUDA_VERSION=13.0.10 B
14RUN |1 TARGETARCH=arm64 /bin/sh -c95.11 MB
15RUN |1 TARGETARCH=arm64 /bin/sh -c185 B
16ENV PATH=/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
17ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/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_PRERELEASE=false0 B
22ARG CUDA_VERSION=13.0.10 B
23ARG PYTHON_VERSION=3.120 B
24ARG NCCL_VERSION=2.30.70 B
25ARG DEADSNAKES_MIRROR_URL0 B
26ARG DEADSNAKES_GPGKEY_URL0 B
27ARG GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py0 B
28ENV CUDA_VERSION=13.0.10 B
29ENV DEBIAN_FRONTEND=noninteractive0 B
30WORKDIR /vllm-workspace104 B
31RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12229 B
32RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12216.03 MB
33RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.121.45 GB
34RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.1232 B
35RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.1237.82 MB
36ENV UV_HTTP_TIMEOUT=5000 B
37ENV UV_INDEX_STRATEGY=unsafe-best-match0 B
38ENV UV_LINK_MODE=copy0 B
39ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python0 B
40ENV UV_CACHE_DIR=/opt/uv/cache0 B
41RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12150 B
42RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12164 B
43ENV UV_OVERRIDE=/etc/uv-overrides.txt0 B
44RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.123.4 KB
45COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit2.03 KB
46RUN |7 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.1232 B
47ENV VLLM_ENABLE_CUDA_COMPATIBILITY=00 B
48ARG PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl0 B
49COPY requirements/common.txt /tmp/common.txt # buildkit1.63 KB
50COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit875 B
51RUN |8 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.124.85 GB
52ARG FLASHINFER_VERSION=0.6.180 B
53RUN |9 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.121.05 GB
54ARG GDRCOPY_CUDA_VERSION=12.80 B
55ARG GDRCOPY_OS_VERSION=Ubuntu24_040 B
56ARG TARGETPLATFORM=linux/arm640 B
57COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit897 B
58RUN |12 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12315.74 KB
59ARG TIMM_VERSION=>=1.0.170 B
60ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.70 B
61RUN |14 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12271.11 MB
62ARG PIP_INDEX_URL UV_INDEX_URL=0 B
63ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL=0 B
64ARG PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl0 B
65ARG PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled0 B
66ARG PYTORCH_NIGHTLY0 B
67COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit201 B
68COPY /workspace/dist/wheel.sha256 /tmp/vllm-wheel.sha256 # buildkit239 B
69RUN |21 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12497.11 MB
70RUN |21 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.1232 B
71ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
72COPY /tmp/ep_kernels_workspace/dist/wheels.sha256 /tmp/ep-kernels-wheels.sha256 # buildkit253 B
73RUN |21 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.1213.65 MB
74ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
75COPY examples examples # buildkit268.21 KB
76COPY benchmarks benchmarks # buildkit284.95 KB
77COPY ./vllm/collect_env.py . # buildkit9.61 KB
78ARG TARGETPLATFORM=linux/arm640 B
79ARG INSTALL_KV_CONNECTORS=true0 B
80ARG INSTALL_RUBIN_PRERELEASE=false0 B
81ARG TRITON_INSTALL_FROM_SOURCE_REPO=0 B
82ARG CUDA_VERSION=13.0.10 B
83ARG VLLM_BUILD_COMMIT0 B
84ARG VLLM_BUILD_PIPELINE0 B
85ARG VLLM_BUILD_URL0 B
86ARG VLLM_IMAGE_TAG0 B
87ARG PIP_INDEX_URL UV_INDEX_URL=0 B
88ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL=0 B
89ENV UV_HTTP_TIMEOUT=5000 B
90ARG torch_cuda_arch_list=8.7 8.9 9.0 10.0+PTX0 B
91ENV TORCH_CUDA_ARCH_LIST=8.7 8.9 9.0 10.0+PTX0 B
92RUN |28 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12526.96 MB
93RUN |28 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.1232 B
94COPY /tmp/triton_from_source_workspace/dist/wheel.sha256 /tmp/triton-source-wheel.sha256 # buildkit212 B
95COPY tools/install_triton_from_source.sh /tmp/install_triton_from_source.sh # buildkit824 B
96RUN |28 INSTALL_RUBIN_PRERELEASE=false CUDA_VERSION=13.0.1 PYTHON_VERSION=3.12174 B
97ENV VLLM_USAGE_SOURCE=production-docker-image0 B
98ENV VLLM_BUILD_COMMIT=unknown VLLM_BUILD_PIPELINE=local VLLM_BUILD_URL= VLLM_IMAGE_TAG=local/vllm-openai:dev0 B
99LABEL 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
100ENTRYPOINT ["vllm" "serve"]0 B

Command

 ARG RELEASE