logo
docker.io/gpustack/runner:cuda13.0-vllm0.33.0
linux/amd64 docker.io firefly 贡献

docker.io/gpustack/runner是一个用于在GPU环境中执行任务的容器镜像,通常集成了GPU驱动、运行时依赖及相关工具组件,支持在容器化场景下高效利用GPU资源运行计算任务(如机器学习训练、推理等GPU加速应用)。

10
浏览次数
27.47GB
镜像大小
国内镜像
swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0
源镜像
docker.io/gpustack/runner:cuda13.0-vllm0.33.0
镜像ID
sha256:c139d638fa4bff3513333b742d7b5baf6267a04454f3e20e2310185e28241387
镜像 TAG
cuda13.0-vllm0.33.0
镜像大小
27.47GB
平台架构
linux/amd64
镜像源
docker.io
CMD
启动入口
tini --
工作目录
/
OS/平台
linux/amd64
镜像创建
2026-11-04T11:47:07.278969633Z
同步时间
2026-11-19 00:44
浏览量
10 次
贡献者
avatar firefly
⚙️ 环境变量 46
KeyValue
PATH=/usr/local/cuda/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin 0
NVARCH=x86_64 1
NVIDIA_REQUIRE_CUDA=cuda>=13.0 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<536 brand=nvidia,driver>=535,driver<536 brand=quadro,driver>=535,driver<536 brand=quadrortx,driver>=535,driver<536 brand=nvidiartx,driver>=535,driver<536 brand=vapps,driver>=535,driver<536 brand=vpc,driver>=535,driver<536 brand=vcs,driver>=535,driver<536 brand=vws,driver>=535,driver<536 brand=cloudgaming,driver>=535,driver<536 brand=unknown,driver>=550,driver<551 brand=grid,driver>=550,driver<551 brand=tesla,driver>=550,driver<551 brand=nvidia,driver>=550,driver<551 brand=quadro,driver>=550,driver<551 brand=quadrortx,driver>=550,driver<551 brand=nvidiartx,driver>=550,driver<551 brand=vapps,driver>=550,driver<551 brand=vpc,driver>=550,driver<551 brand=vcs,driver>=550,driver<551 brand=vws,driver>=550,driver<551 brand=cloudgaming,driver>=550,driver<551 brand=unknown,driver>=565,driver<566 brand=grid,driver>=565,driver<566 brand=tesla,driver>=565,driver<566 brand=nvidia,driver>=565,driver<566 brand=quadro,driver>=565,driver<566 brand=quadrortx,driver>=565,driver<566 brand=nvidiartx,driver>=565,driver<566 brand=vapps,driver>=565,driver<566 brand=vpc,driver>=565,driver<566 brand=vcs,driver>=565,driver<566 brand=vws,driver>=565,driver<566 brand=cloudgaming,driver>=565,driver<566 brand=unknown,driver>=570,driver<571 brand=grid,driver>=570,driver<571 brand=tesla,driver>=570,driver<571 brand=nvidia,driver>=570,driver<571 brand=quadro,driver>=570,driver<571 brand=quadrortx,driver>=570,driver<571 brand=nvidiartx,driver>=570,driver<571 brand=vapps,driver>=570,driver<571 brand=vpc,driver>=570,driver<571 brand=vcs,driver>=570,driver<571 brand=vws,driver>=570,driver<571 brand=cloudgaming,driver>=570,driver<571 brand=unknown,driver>=575,driver<576 brand=grid,driver>=575,driver<576 brand=tesla,driver>=575,driver<576 brand=nvidia,driver>=575,driver<576 brand=quadro,driver>=575,driver<576 brand=quadrortx,driver>=575,driver<576 brand=nvidiartx,driver>=575,driver<576 brand=vapps,driver>=575,driver<576 brand=vpc,driver>=575,driver<576 brand=vcs,driver>=575,driver<576 brand=vws,driver>=575,driver<576 brand=cloudgaming,driver>=575,driver<576 2
NV_CUDA_CUDART_VERSION=13.0.96-1 3
CUDA_VERSION=13.0.3 4
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/lib64 5
NVIDIA_VISIBLE_DEVICES=all 6
NVIDIA_DRIVER_CAPABILITIES=compute,utility 7
DEBIAN_FRONTEND=noninteractive 8
CUDA_HOME=/usr/local/cuda 9
CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu 10
UV_HTTP_TIMEOUT=500 11
UV_INDEX_STRATEGY=unsafe-best-match 12
UV_LINK_MODE=copy 13
UV_PYTHON_INSTALL_DIR=/opt/uv/python 14
UV_CACHE_DIR=/opt/uv/cache 15
UV_COMPILE_BYTECODE=1 16
UV_OVERRIDE=/etc/uv-overrides.txt 17
VLLM_ENABLE_CUDA_COMPATIBILITY=0 18
TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 8.9 9.0 10.0 12.0 19
VLLM_USAGE_SOURCE=production-docker-image 20
VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5 21
VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b 22
VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673 23
VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0 24
LANG=en_US.UTF-8 25
LANGUAGE=en_US:en 26
LC_ALL=en_US.UTF-8 27
PYTHON_VERSION=3.12 28
PIP_NO_CACHE_DIR=1 29
PIP_DISABLE_PIP_VERSION_CHECK=1 30
PIP_ROOT_USER_ACTION=ignore 31
CUDA_ARCHS= 32
UV_SYSTEM_PYTHON=1 33
UV_PRERELEASE=allow 34
UV_SKIP_WHEEL_FILENAME_CHECK=1 35
UV_NO_CACHE=1 36
VLLM_VERSION=0.33.0 37
VLLM_TORCH_VERSION=2.14.0 38
VLLM_TORCH_CUDA_VERSION=13.0.3 39
VLLM_LMCACHE_VERSION=0.5.6 40
RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES=1 41
SAFETENSORS_FAST_GPU=1 42
RUNAI_STREAMER_MEMORY_LIMIT=0 43
RUNAI_STREAMER_LOG_TO_STDERR=1 44
RUNAI_STREAMER_LOG_LEVEL=INFO 45
🏷️ 镜像标签 10
KeyValue
4ba787ec9959e566d3b587a796dff07673324da5 ai.vllm.build.commit
019d130e-464e-4ff7-b84b-492992c0c06b ai.vllm.build.pipeline
https://buildkite.com/vllm/release-v2/builds/9673 ai.vllm.build.url
vllm/vllm-openai:v0.33.0 ai.vllm.image.tag
NVIDIA CORPORATION <cudatools@nvidia.com> maintainer
ubuntu org.opencontainers.image.ref.name
4ba787ec9959e566d3b587a796dff07673324da5 org.opencontainers.image.revision
https://github.com/vllm-project/vllm org.opencontainers.image.source
https://buildkite.com/vllm/release-v2/builds/9673 org.opencontainers.image.url
vllm/vllm-openai:v0.33.0 org.opencontainers.image.version

Docker拉取命令

docker pull swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0
docker tag  swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0  docker.io/gpustack/runner:cuda13.0-vllm0.33.0

Containerd拉取命令

ctr images pull swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0
ctr images tag  swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0  docker.io/gpustack/runner:cuda13.0-vllm0.33.0

Shell快速替换命令

sed -i 's#gpustack/runner:cuda13.0-vllm0.33.0#swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0#' deployment.yaml

Ansible快速分发-Docker

#ansible k8s -m shell -a 'docker pull swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0 && docker tag  swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0  docker.io/gpustack/runner:cuda13.0-vllm0.33.0'

Ansible快速分发-Containerd

#ansible k8s -m shell -a 'ctr images pull swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0 && ctr images tag  swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0  docker.io/gpustack/runner:cuda13.0-vllm0.33.0'

镜像构建历史


# 2026-11-04 19:47:07  0.00B 配置容器启动时运行的命令
ENTRYPOINT ["tini" "--"]
                        
# 2026-11-04 19:47:07  0.00B 设置工作目录为/
WORKDIR /
                        
# 2026-11-04 19:47:07  0.00B 设置环境变量 RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES SAFETENSORS_FAST_GPU RUNAI_STREAMER_MEMORY_LIMIT RUNAI_STREAMER_LOG_TO_STDERR RUNAI_STREAMER_LOG_LEVEL
ENV RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES=1 SAFETENSORS_FAST_GPU=1 RUNAI_STREAMER_MEMORY_LIMIT=0 RUNAI_STREAMER_LOG_TO_STDERR=1 RUNAI_STREAMER_LOG_LEVEL=INFO
                        
# 2026-11-04 19:47:07  0.00B 执行命令并创建新的镜像层
RUN |12 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 VLLM_LMCACHE_VERSION=0.5.6 VLLM_TORCHCODEC_VERSION= /bin/bash -eo pipefail -c     # TorchCodec

    if [[ -z "${VLLM_TORCHCODEC_VERSION}" ]]; then
        echo "Keeping the base image's torchcodec..."
        exit 0
    fi

    # Replace the base's torchcodec with the CPU build of the same version ——
    # see the note on VLLM_TORCHCODEC_VERSION. --no-deps keeps uv from pulling
    # a CPU torch over the base's CUDA build.
    uv pip install --no-deps \
        "torchcodec==${VLLM_TORCHCODEC_VERSION}+cpu" \
        --index-url https://download.pytorch.org/whl/cpu

    # Fail the build on the import path that crashes the stock cu129 image
    # (vllm-project/vllm#60406); needs no GPU.
    python -c "import torch, torchcodec.decoders; print('torch', torch.__version__, 'cuda', torch.version.cuda)"

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:47:07  0.00B 定义构建参数
ARG VLLM_TORCHCODEC_VERSION
                        
# 2026-11-04 19:47:07  63.12MB 执行命令并创建新的镜像层
RUN |11 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 VLLM_LMCACHE_VERSION=0.5.6 /bin/bash -eo pipefail -c     # Dependencies

    # Install
    cat <<EOT >/tmp/requirements.txt
# audio extras
librosa
soundfile
mistral_common[audio]

# tokenizer extras
fastokens==0.2.0
EOT
    uv pip install \
        -r /tmp/requirements.txt

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:47:05  112.15KB 执行命令并创建新的镜像层
RUN |11 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 VLLM_LMCACHE_VERSION=0.5.6 /bin/bash -eo pipefail -c     # Patch

    tree -hs /workspace/patches
    pushd $(pip show vllm | grep Location: | cut -d" " -f 2) \
        && for f in /workspace/patches/vllm/*.patch; do patch -p1 < "$f" || exit 1; done
    if pip show vllm_omni > /dev/null 2>&1; then \
        pushd $(pip show vllm_omni | grep Location: | cut -d" " -f 2) \
            && patch -p1 < /workspace/patches/vllm_omni/*.patch; \
    fi
 # buildkit
                        
# 2026-11-04 19:47:04  0.00B 执行命令并创建新的镜像层
RUN |11 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 VLLM_LMCACHE_VERSION=0.5.6 /bin/bash -eo pipefail -c     # Postprocess

    # Review
    uv pip tree \
        --package vllm \
        --package vllm-omni \
        --package flashinfer-python \
        --package torch \
        --package triton \
        --package deep-gemm \
        --package deep-ep \
        --package lmcache \
        --package mooncake-transfer-engine \
        --package ray
 # buildkit
                        
# 2026-11-04 19:47:04  195.47MB 执行命令并创建新的镜像层
RUN |11 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 VLLM_LMCACHE_VERSION=0.5.6 /bin/bash -eo pipefail -c     # CuPy
    # ray[cgraph] pulls cupy-cuda12x while a cu13 base already ships cupy-cuda13x; both share
    # the same cupy/ tree, so uninstalling one breaks the other. Purge both, then reinstall
    # only the CUDA-matching variant at the version vLLM pinned in the base, so cu129 also
    # ends on that pinned version instead of ray's unconstrained cupy-cuda12x.
    # Our own lmcache is a third source but agrees with this reinstall: its build stage sets
    # LMCACHE_CUDA_MAJOR from the same CUDA major used here.

    IFS="." read -r CUDA_MAJOR CUDA_MINOR CUDA_PATCH <<< "${CUDA_VERSION}"
    PLATLIB="$(python -c 'import sysconfig; print(sysconfig.get_paths()["platlib"])')"
    CUPY="cupy-cuda${CUDA_MAJOR}x"
    # The base ships the variant matching its own CUDA major (cupy-cuda13x on cu13,
    # cupy-cuda12x on cu12), so read the pinned version from that same variant.
    CUPY_VER="$(uv pip show "${CUPY}" 2>/dev/null | awk '/^Version:/{print $2}' || true)"
    uv pip uninstall cupy-cuda12x cupy-cuda13x || true
    rm -rf "${PLATLIB}"/cupy "${PLATLIB}"/cupyx "${PLATLIB}"/cupy_backends
    # cupy-cuda12x and cupy-cuda13x share the same version number, so reuse the base's
    # pinned version and install only the variant matching the CUDA major we build for.
    # Fail loudly if the base no longer ships cupy, rather than installing an unpinned latest.
    if [[ -z "${CUPY_VER}" ]]; then
        echo "${CUPY} not found in base image; refusing to install unpinned cupy" >&2
        exit 1
    fi
    uv pip install "${CUPY}==${CUPY_VER}"

    # Review
    uv pip show "${CUPY}"

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:47:02  84.06MB 执行命令并创建新的镜像层
RUN |11 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 VLLM_LMCACHE_VERSION=0.5.6 /bin/bash -eo pipefail -c     # LMCache

    # Install
    ## Replaces the base's own lmcache —— see the note on VLLM_LMCACHE_VERSION. Must run before
    ## the CuPy dedup below, since lmcache declares a CUDA-major-specific cupy variant.
    uv pip install --no-build-isolation \
        /lmcache/workspace/*.whl

    # Review
    uv pip show lmcache
    ## Fail here on a torch ABI mismatch: lmcache itself only WARNs and falls back to its slow
    ## torch baseline. RTLD_NOW resolves all symbols without module init, so this needs no GPU.
    cat <<EOT >/tmp/check_lmcache.py
import ctypes, glob, os, sysconfig, torch
platlib = sysconfig.get_paths()["platlib"]
sos = glob.glob(platlib + "/lmcache/*.so")
assert sos, "no lmcache extension found under " + platlib
for so in sos:
    ctypes.CDLL(so, mode=os.RTLD_NOW)
print("lmcache resolved against torch " + torch.__version__ + ": " + ", ".join(sos))
EOT
    python /tmp/check_lmcache.py

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:42:47  0.00B 设置环境变量 VLLM_LMCACHE_VERSION
ENV VLLM_LMCACHE_VERSION=0.5.6
                        
# 2026-11-04 19:42:47  0.00B 定义构建参数
ARG VLLM_LMCACHE_VERSION=0.5.6
                        
# 2026-11-04 19:42:47  1.07GB 执行命令并创建新的镜像层
RUN |10 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 /bin/bash -eo pipefail -c     # Omni

    if [[ ! -d /omni/workspace ]]; then
        echo "Skipping vLLM Omni installation for ${TARGETARCH}..."
        exit 0
    fi

    # Install
    uv pip install --no-build-isolation \
        /omni/workspace/*.whl

    # Dependencies
    cat <<EOT >/tmp/requirements.txt
onnxruntime
sox
EOT
    uv pip install \
        -r /tmp/requirements.txt

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:42:01  522.61MB 执行命令并创建新的镜像层
RUN |10 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 /bin/bash -eo pipefail -c     # Ray

    # Install Ray Client and Default.
    # Pinned to 2.54.0: ray 2.55 (ray-project/ray#61029) breaks distributed
    # ray.init when the driver has no co-located raylet (gpustack split containers).
    RAY_VERSION=2.54.0
    cat <<EOT >/tmp/requirements.txt
ray[cgraph]==${RAY_VERSION}
ray[client]==${RAY_VERSION}
ray[default]==${RAY_VERSION}
EOT
    uv pip install \
        -r /tmp/requirements.txt

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:41:55  72.76MB 执行命令并创建新的镜像层
RUN |10 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 CUDA_VERSION=13.0.3 CUDA_ARCHS= VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3 VLLM_ROUTER_VERSION=0.1.17 /bin/bash -eo pipefail -c     # vLLM Router
    #
    # The PD router GPUStack renders for a disaggregated group. Without it in
    # the image there is no router to run, and the group either falls back to
    # an engine example script (no metrics, no circuit breaker, no health
    # check) or does not start at all.
    uv pip install \
        vllm-router==${VLLM_ROUTER_VERSION}

    # Fail the build rather than ship an image whose router is missing: the
    # alternative surfaces as a container that exits at deploy time, one
    # layer away from anything that explains it.
    command -v vllm-router >/dev/null

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG VLLM_ROUTER_VERSION=0.1.17
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG TARGETARCH=amd64
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG TARGETOS=linux
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG TARGETPLATFORM=linux/amd64
                        
# 2026-11-04 19:41:54  0.00B 
SHELL [/bin/bash -eo pipefail -c]
                        
# 2026-11-04 19:41:54  0.00B 设置环境变量 VLLM_VERSION VLLM_TORCH_VERSION VLLM_TORCH_CUDA_VERSION
ENV VLLM_VERSION=0.33.0 VLLM_TORCH_VERSION=2.14.0 VLLM_TORCH_CUDA_VERSION=13.0.3
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG VLLM_TORCH_CUDA_VERSION=13.0.3
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG VLLM_TORCH_VERSION=2.14.0
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG VLLM_VERSION=0.33.0
                        
# 2026-11-04 19:41:54  0.00B 设置环境变量 UV_SYSTEM_PYTHON UV_PRERELEASE UV_SKIP_WHEEL_FILENAME_CHECK UV_NO_CACHE UV_HTTP_TIMEOUT UV_INDEX_STRATEGY
ENV UV_SYSTEM_PYTHON=1 UV_PRERELEASE=allow UV_SKIP_WHEEL_FILENAME_CHECK=1 UV_NO_CACHE=1 UV_HTTP_TIMEOUT=500 UV_INDEX_STRATEGY=unsafe-best-match
                        
# 2026-11-04 19:41:54  0.00B 设置环境变量 CUDA_HOME CUDA_VERSION CUDA_ARCHS
ENV CUDA_HOME=/usr/local/cuda CUDA_VERSION=13.0.3 CUDA_ARCHS=
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG CUDA_ARCHS
                        
# 2026-11-04 19:41:54  0.00B 定义构建参数
ARG CUDA_VERSION=13.0.3
                        
# 2026-11-04 19:41:54  84.85MB 执行命令并创建新的镜像层
RUN |4 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 /bin/bash -eo pipefail -c     # Buildkit

    cat <<EOT >/tmp/requirements.txt
build
cmake<4
wheel
pybind11<3
Cython
EOT
    pip install -r /tmp/requirements.txt

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/*
 # buildkit
                        
# 2026-11-04 19:41:51  0.00B 设置环境变量 PIP_NO_CACHE_DIR PIP_DISABLE_PIP_VERSION_CHECK PIP_ROOT_USER_ACTION
ENV PIP_NO_CACHE_DIR=1 PIP_DISABLE_PIP_VERSION_CHECK=1 PIP_ROOT_USER_ACTION=ignore
                        
# 2026-11-04 19:41:51  16.81KB 执行命令并创建新的镜像层
RUN |4 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 PYTHON_VERSION=3.12 /bin/bash -eo pipefail -c     # Python

    # Update alternatives
    if ! command -v python >/dev/null 2>&1 \
        && command -v python${PYTHON_VERSION} >/dev/null 2>&1; then
        update-alternatives --install /usr/bin/python python $(which python${PYTHON_VERSION}) 1
    fi
 # buildkit
                        
# 2026-11-04 19:41:51  0.00B 设置环境变量 PYTHON_VERSION
ENV PYTHON_VERSION=3.12
                        
# 2026-11-04 19:41:51  0.00B 定义构建参数
ARG PYTHON_VERSION=3.12
                        
# 2026-11-04 19:41:51  219.36MB 执行命令并创建新的镜像层
RUN |3 TARGETPLATFORM=linux/amd64 TARGETOS=linux TARGETARCH=amd64 /bin/bash -eo pipefail -c     # Tools

    # Refresh
    apt-get update -y && apt-get install -y --no-install-recommends \
        software-properties-common apt-transport-https \
        ca-certificates gnupg2 lsb-release gnupg-agent \
      && apt-get update -y

    # Install
    apt-get install -y --no-install-recommends \
        ca-certificates build-essential binutils bash openssl \
        curl wget aria2 \
        git git-lfs \
        unzip xz-utils \
        tzdata locales \
        iproute2 iputils-ping ifstat net-tools dnsutils pciutils ipmitool \
        rdma-core rdmacm-utils infiniband-diags \
        procps sysstat htop \
        tini vim jq bc tree

    # Update locale
    localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8

    # Update timezone
    rm -f /etc/localtime \
        && ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime \
        && echo "Asia/Shanghai" > /etc/timezone \
        && dpkg-reconfigure --frontend noninteractive tzdata

    # Cleanup
    rm -rf /var/tmp/* \
        && rm -rf /tmp/* \
        && rm -rf /var/cache/apt
 # buildkit
                        
# 2026-11-04 19:41:51  0.00B 设置环境变量 DEBIAN_FRONTEND LANG LANGUAGE LC_ALL
ENV DEBIAN_FRONTEND=noninteractive LANG=en_US.UTF-8 LANGUAGE=en_US:en LC_ALL=en_US.UTF-8
                        
# 2026-11-04 19:41:51  0.00B 定义构建参数
ARG TARGETARCH=amd64
                        
# 2026-11-04 19:41:51  0.00B 定义构建参数
ARG TARGETOS=linux
                        
# 2026-11-04 19:41:51  0.00B 定义构建参数
ARG TARGETPLATFORM=linux/amd64
                        
# 2026-11-04 19:41:51  0.00B 
SHELL [/bin/bash -eo pipefail -c]
                        
# 2026-11-01 17:45:11  0.00B 配置容器启动时运行的命令
ENTRYPOINT ["vllm" "serve"]
                        
# 2026-11-01 17:45:11  0.00B 添加元数据标签
LABEL org.opencontainers.image.source=https://github.com/vllm-project/vllm org.opencontainers.image.revision=4ba787ec9959e566d3b587a796dff07673324da5 org.opencontainers.image.version=vllm/vllm-openai:v0.33.0 org.opencontainers.image.url=https://buildkite.com/vllm/release-v2/builds/9673 ai.vllm.build.commit=4ba787ec9959e566d3b587a796dff07673324da5 ai.vllm.build.pipeline=019d130e-464e-4ff7-b84b-492992c0c06b ai.vllm.build.url=https://buildkite.com/vllm/release-v2/builds/9673 ai.vllm.image.tag=vllm/vllm-openai:v0.33.0
                        
# 2026-11-01 17:45:11  0.00B 设置环境变量 VLLM_BUILD_COMMIT VLLM_BUILD_PIPELINE VLLM_BUILD_URL VLLM_IMAGE_TAG
ENV VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673 VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0
                        
# 2026-11-01 17:45:11  0.00B 设置环境变量 VLLM_USAGE_SOURCE
ENV VLLM_USAGE_SOURCE=production-docker-image
                        
# 2026-11-01 17:45:11  0.00B 执行命令并创建新的镜像层
RUN |15 TARGETPLATFORM=linux/amd64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false TRITON_INSTALL_FROM_SOURCE_REPO= CUDA_VERSION=13.0.2 VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673 VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= torch_cuda_arch_list=7.5 8.0 8.6 8.9 9.0 10.0 12.0 BUILD_NIXL=false /bin/sh -c set -eu;     if [ "${BUILD_NIXL}" = "true" ]; then         ls /tmp/nixl-dist/*.whl;         uv pip uninstall --system nixl nixl-cu12 nixl-cu13;         uv pip install --system --no-deps /tmp/nixl-dist/*.whl;         python3 -c "import glob, importlib.util as u, os, torch;             spec = u.find_spec('nixl_ep_cu' + os.environ['CUDA_VERSION'].split('.')[0]);             mm = ''.join(torch.__version__.split('.')[:2]);             ext = glob.glob(os.path.join(spec.submodule_search_locations[0], 'nixl_ep_cpp_torch' + mm + '.*')) if spec else [];             assert ext, 'NIXL EP has no build for torch ' + torch.__version__;             print('NIXL EP extension:', ext[0])";     fi # buildkit
                        
# 2026-11-01 17:45:10  0.00B 定义构建参数
ARG BUILD_NIXL
                        
# 2026-11-01 17:45:10  0.00B 执行命令并创建新的镜像层
RUN |14 TARGETPLATFORM=linux/amd64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false TRITON_INSTALL_FROM_SOURCE_REPO= CUDA_VERSION=13.0.2 VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673 VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= torch_cuda_arch_list=7.5 8.0 8.6 8.9 9.0 10.0 12.0 /bin/sh -c if [ -n "${TRITON_INSTALL_FROM_SOURCE_REPO}" ]; then         CUDA_MAJOR_MINOR=$(echo "${CUDA_VERSION}" | cut -d. -f1,2);         if [ "${INSTALL_RUBIN_PRERELEASE}" = "true" ]; then             case "${CUDA_MAJOR_MINOR}" in                 13.4|13.5)                     TRITON_SYSTEM_PTXAS_PATH=/usr/local/cuda/bin/ptxas                     TRITON_EXPECTED_PTXAS_VERSION="${CUDA_MAJOR_MINOR}"                         bash /tmp/install_triton_from_source.sh                             /tmp/triton-source-dist                             /tmp/triton-source-wheel.sha256                     ;;                 *)                     echo "Unsupported Rubin CUDA version: ${CUDA_MAJOR_MINOR}" >&2;                     exit 1                     ;;             esac;         else             bash /tmp/install_triton_from_source.sh                 /tmp/triton-source-dist                 /tmp/triton-source-wheel.sha256;         fi;     fi     && rm -f         /tmp/install_triton_from_source.sh         /tmp/triton-source-wheel.sha256 # buildkit
                        
# 2026-11-01 17:45:10  1.51KB 复制新文件或目录到容器中
COPY tools/install_triton_from_source.sh /tmp/install_triton_from_source.sh # buildkit
                        
# 2026-11-01 17:45:10  78.00B 复制新文件或目录到容器中
COPY /tmp/triton_from_source_workspace/dist/wheel.sha256 /tmp/triton-source-wheel.sha256 # buildkit
                        
# 2026-11-01 17:45:09  0.00B 执行命令并创建新的镜像层
RUN |14 TARGETPLATFORM=linux/amd64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false TRITON_INSTALL_FROM_SOURCE_REPO= CUDA_VERSION=13.0.2 VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673 VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= torch_cuda_arch_list=7.5 8.0 8.6 8.9 9.0 10.0 12.0 /bin/sh -c if [ "${INSTALL_RUBIN_PRERELEASE}" = "true" ]; then         echo "Installing Rubin prerelease requirements..."         && uv pip install --system             --prerelease=allow --index-strategy unsafe-first-match             -r /tmp/vllm-requirements/rubin-prerelease.txt;     fi # buildkit
                        
# 2026-11-01 17:45:09  1.25GB 执行命令并创建新的镜像层
RUN |14 TARGETPLATFORM=linux/amd64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false TRITON_INSTALL_FROM_SOURCE_REPO= CUDA_VERSION=13.0.2 VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673 VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= torch_cuda_arch_list=7.5 8.0 8.6 8.9 9.0 10.0 12.0 /bin/sh -c set -eu;     CUDA_MAJOR="${CUDA_VERSION%%.*}";     CUDA_VERSION_DASH=$(echo $CUDA_VERSION | cut -d. -f1,2 | tr '.' '-');     CUDA_HOME=/usr/local/cuda;     BUILD_PKGS="libcusparse-dev-${CUDA_VERSION_DASH}                 libcublas-dev-${CUDA_VERSION_DASH}                 libcusolver-dev-${CUDA_VERSION_DASH}";     if [ "$INSTALL_KV_CONNECTORS" = "true" ]; then         KV_REQUIREMENTS=/tmp/vllm-requirements/kv_connectors.txt;         if [ "${CUDA_MAJOR}" = "12" ]; then             KV_REQUIREMENTS=/tmp/vllm-requirements/kv_connectors_cu12.txt;         fi;         uv pip freeze --system > /tmp/kv-installed.txt;         grep -E '^(torch|torchaudio|torchvision|triton|cuda-bindings|cuda-python|cuda-toolkit)=='             /tmp/kv-installed.txt > /tmp/kv-constraints.txt;         uv pip install --system -c /tmp/kv-constraints.txt -r "${KV_REQUIREMENTS}" --no-build || (             apt-get update -y &&             apt-get install -y --no-install-recommends --allow-change-held-packages ${BUILD_PKGS} &&             uv pip install --system -c /tmp/kv-constraints.txt -r "${KV_REQUIREMENTS}" --no-build-isolation &&             apt-get purge -y ${BUILD_PKGS} &&             rm -rf /var/lib/apt/lists/*         );         MOONCAKE_VERSION=$(python3 -c "import importlib.metadata as m; print(m.version('mooncake-transfer-engine'))" 2>/dev/null || true) &&         if [ "${CUDA_MAJOR}" = "13" ] && [ -n "${MOONCAKE_VERSION}" ]; then             uv pip uninstall --system mooncake-transfer-engine 2>/dev/null || true;             uv pip install --system -c /tmp/kv-constraints.txt "mooncake-transfer-engine-cuda13==${MOONCAKE_VERSION}";         fi;         rm -f /tmp/kv-installed.txt /tmp/kv-constraints.txt;     fi # buildkit
                        
# 2026-11-01 17:45:00  0.00B 设置环境变量 TORCH_CUDA_ARCH_LIST
ENV TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 8.9 9.0 10.0 12.0
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG torch_cuda_arch_list=7.5 8.0 8.6 8.9 9.0 10.0 11.0 12.0
                        
# 2026-11-01 17:45:00  0.00B 设置环境变量 UV_HTTP_TIMEOUT
ENV UV_HTTP_TIMEOUT=500
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG PIP_INDEX_URL UV_INDEX_URL
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG VLLM_IMAGE_TAG
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG VLLM_BUILD_URL
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG VLLM_BUILD_PIPELINE
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG VLLM_BUILD_COMMIT
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG CUDA_VERSION
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG TRITON_INSTALL_FROM_SOURCE_REPO
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG INSTALL_RUBIN_PRERELEASE
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG INSTALL_KV_CONNECTORS=false
                        
# 2026-11-01 17:45:00  0.00B 定义构建参数
ARG TARGETPLATFORM
                        
# 2026-11-01 17:45:00  35.04KB 复制新文件或目录到容器中
COPY ./vllm/collect_env.py . # buildkit
                        
# 2026-11-01 17:45:00  1.33MB 复制新文件或目录到容器中
COPY benchmarks benchmarks # buildkit
                        
# 2026-11-01 17:45:00  1.51MB 复制新文件或目录到容器中
COPY examples examples # buildkit
                        
# 2026-11-01 17:45:00  0.00B 设置环境变量 LD_LIBRARY_PATH
ENV 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/lib64
                        
# 2026-11-01 17:45:00  43.48MB 执行命令并创建新的镜像层
RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled PYTORCH_NIGHTLY= /bin/sh -c uv pip install --system ep_kernels/dist/*.whl --verbose         --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.') # buildkit
                        
# 2026-11-01 17:44:59  146.00B 复制新文件或目录到容器中
COPY /tmp/ep_kernels_workspace/dist/wheels.sha256 /tmp/ep-kernels-wheels.sha256 # buildkit
                        
# 2026-11-01 17:44:58  0.00B 设置环境变量 LD_LIBRARY_PATH
ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib64
                        
# 2026-11-01 17:44:58  0.00B 执行命令并创建新的镜像层
RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled PYTORCH_NIGHTLY= /bin/sh -c . /etc/environment && uv pip list # buildkit
                        
# 2026-11-01 17:44:58  52.00B 执行命令并创建新的镜像层
RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled PYTORCH_NIGHTLY= /bin/sh -c ln -s "$(python3 -c 'import sysconfig; print(sysconfig.get_path("platlib"))')/vllm/vllm-rs" /usr/local/bin/vllm-rs     && vllm-rs --version # buildkit
                        
# 2026-11-01 17:44:57  892.09MB 执行命令并创建新的镜像层
RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL= UV_EXTRA_INDEX_URL= PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl PIP_KEYRING_PROVIDER=disabled UV_KEYRING_PROVIDER=disabled PYTORCH_NIGHTLY= /bin/sh -c if [ "${PYTORCH_NIGHTLY}" = "1" ]; then         echo "Installing torch nightly..."         && uv pip install --system $(cat torch_lib_versions.txt | xargs) --pre         --index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/nightly/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')         && echo "Installing vLLM..."         && uv pip install --system dist/*.whl --verbose         --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/nightly/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.');     else         echo "Installing vLLM..."         && uv pip install --system dist/*.whl --verbose         --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.');     fi # buildkit
                        
# 2026-11-01 17:44:51  110.00B 复制新文件或目录到容器中
COPY /workspace/dist/wheel.sha256 /tmp/vllm-wheel.sha256 # buildkit
                        
# 2026-11-01 17:30:34  71.00B 复制新文件或目录到容器中
COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit
                        
# 2026-11-01 17:30:34  0.00B 定义构建参数
ARG PYTORCH_NIGHTLY
                        
# 2026-11-01 17:30:34  0.00B 定义构建参数
ARG PIP_KEYRING_PROVIDER UV_KEYRING_PROVIDER
                        
# 2026-11-01 17:30:34  0.00B 定义构建参数
ARG PYTORCH_CUDA_INDEX_BASE_URL
                        
# 2026-11-01 17:30:34  0.00B 定义构建参数
ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL
                        
# 2026-11-01 17:30:34  0.00B 定义构建参数
ARG PIP_INDEX_URL UV_INDEX_URL
                        
# 2026-11-01 17:30:34  178.97MB 执行命令并创建新的镜像层
RUN |16 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl GDRCOPY_CUDA_VERSION=12.8 GDRCOPY_OS_VERSION=Ubuntu24_04 TARGETPLATFORM=linux/amd64 TIMM_VERSION=>=1.0.17 RUNAI_MODEL_STREAMER_VERSION=>=0.15.7 GGUF_VERSION=>=0.19.0 /bin/sh -c uv pip install --system accelerate 'modelscope<1.38'         "timm${TIMM_VERSION}" "runai-model-streamer[s3,gcs,azure]${RUNAI_MODEL_STREAMER_VERSION}"         "gguf${GGUF_VERSION}" # buildkit
                        
# 2026-11-01 17:30:31  0.00B 定义构建参数
ARG GGUF_VERSION=>=0.19.0
                        
# 2026-11-01 17:30:31  0.00B 定义构建参数
ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.7
                        
# 2026-11-01 17:30:31  0.00B 定义构建参数
ARG TIMM_VERSION=>=1.0.17
                        
# 2026-11-01 17:30:31  2.57MB 执行命令并创建新的镜像层
RUN |13 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl GDRCOPY_CUDA_VERSION=12.8 GDRCOPY_OS_VERSION=Ubuntu24_04 TARGETPLATFORM=linux/amd64 /bin/sh -c set -eux;     case "${TARGETPLATFORM}" in       linux/arm64) UUARCH="aarch64" ;;       linux/amd64) UUARCH="x64" ;;       *) echo "Unsupported TARGETPLATFORM: ${TARGETPLATFORM}" >&2; exit 1 ;;     esac;     /tmp/install_gdrcopy.sh "${GDRCOPY_OS_VERSION}" "${GDRCOPY_CUDA_VERSION}" "${UUARCH}" &&     rm /tmp/install_gdrcopy.sh # buildkit
                        
# 2026-11-01 17:30:25  1.44KB 复制新文件或目录到容器中
COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit
                        
# 2026-11-01 17:30:25  0.00B 定义构建参数
ARG TARGETPLATFORM
                        
# 2026-11-01 17:30:25  0.00B 定义构建参数
ARG GDRCOPY_OS_VERSION=Ubuntu24_04
                        
# 2026-11-01 17:30:25  0.00B 定义构建参数
ARG GDRCOPY_CUDA_VERSION=12.8
                        
# 2026-11-01 17:30:25  9.77GB 执行命令并创建新的镜像层
RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl /bin/sh -c PIP_CACHE_DIR=/opt/pip/cache flashinfer download-kernels # buildkit
                        
# 2026-11-01 17:28:56  8.91GB 执行命令并创建新的镜像层
RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py PYTORCH_CUDA_INDEX_BASE_URL=https://download.pytorch.org/whl /bin/sh -c if [ "${INSTALL_RUBIN_PRERELEASE}" = "true" ] &&         [ ! -f /tmp/vllm-requirements/rubin-prerelease.txt ]; then         echo "INSTALL_RUBIN_PRERELEASE=true requires requirements/rubin-prerelease.txt" >&2;         exit 1;     fi &&     if [ "$(echo $CUDA_VERSION | cut -d. -f1)" = "12" ]; then         sed -i 's/^nvidia-cutlass-dsl\[cu13\]/nvidia-cutlass-dsl/' /tmp/requirements-cuda.txt;         sed -i 's/^humming-kernels\[cu13\]/humming-kernels[cu12]/' /tmp/requirements-cuda.txt;     fi &&     if [ "${INSTALL_RUBIN_PRERELEASE}" = "true" ]; then         uv pip install --system             --prerelease=allow --index-strategy unsafe-first-match             -r /tmp/vllm-requirements/rubin-prerelease.txt         && sed -Ei '/^[[:space:]]*(torch|torchvision|torchaudio)[[:space:]]*==/Id'             /tmp/requirements-cuda.txt         && uv pip install --system -r /tmp/requirements-cuda.txt         && uv pip install --system             --prerelease=allow --index-strategy unsafe-first-match             -r /tmp/vllm-requirements/rubin-prerelease.txt;     else         uv pip install --system             -r /tmp/requirements-cuda.txt             --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.');     fi &&     rm /tmp/requirements-cuda.txt /tmp/common.txt # buildkit
                        
# 2026-11-01 17:28:11  1.10KB 复制新文件或目录到容器中
COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit
                        
# 2026-11-01 17:28:11  3.37KB 复制新文件或目录到容器中
COPY requirements/common.txt /tmp/common.txt # buildkit
                        
# 2026-11-01 17:28:11  0.00B 定义构建参数
ARG PYTORCH_CUDA_INDEX_BASE_URL
                        
# 2026-11-01 17:28:11  0.00B 设置环境变量 VLLM_ENABLE_CUDA_COMPATIBILITY
ENV VLLM_ENABLE_CUDA_COMPATIBILITY=0
                        
# 2026-11-01 17:28:11  3.69KB 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c chmod 0755 /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit
                        
# 2026-11-01 17:28:10  3.69KB 复制新文件或目录到容器中
COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit
                        
# 2026-11-01 17:28:10  9.53KB 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c useradd --uid 2000 --gid 0 --create-home --home-dir /home/vllm         --shell /bin/bash vllm     && mkdir -p /home/vllm/.cache /home/vllm/.config     && chown -R 2000:0 /home/vllm     && chmod -R g+rwX /home/vllm     && chgrp 0 /etc/passwd /etc/group     && chmod g=u /etc/passwd /etc/group # buildkit
                        
# 2026-11-01 17:28:09  0.00B 设置环境变量 UV_OVERRIDE
ENV UV_OVERRIDE=/etc/uv-overrides.txt
                        
# 2026-11-01 17:28:09  25.00B 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c CUDA_MAJOR=$(echo $CUDA_VERSION | cut -d. -f1) &&     if [ -n "$NCCL_VERSION" ]; then         echo "nvidia-nccl-cu${CUDA_MAJOR}==${NCCL_VERSION}"             > /etc/uv-overrides.txt;     else         touch /etc/uv-overrides.txt;     fi # buildkit
                        
# 2026-11-01 17:28:09  0.00B 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c mkdir -p "${UV_PYTHON_INSTALL_DIR}" "${UV_CACHE_DIR}"     && chgrp -R 0 /opt/uv     && chmod -R g+rwX,a+rX /opt/uv # buildkit
                        
# 2026-11-01 17:28:08  0.00B 设置环境变量 UV_COMPILE_BYTECODE
ENV UV_COMPILE_BYTECODE=1
                        
# 2026-11-01 17:28:08  0.00B 设置环境变量 UV_CACHE_DIR
ENV UV_CACHE_DIR=/opt/uv/cache
                        
# 2026-11-01 17:28:08  0.00B 设置环境变量 UV_PYTHON_INSTALL_DIR
ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python
                        
# 2026-11-01 17:28:08  0.00B 设置环境变量 UV_LINK_MODE
ENV UV_LINK_MODE=copy
                        
# 2026-11-01 17:28:08  0.00B 设置环境变量 UV_INDEX_STRATEGY
ENV UV_INDEX_STRATEGY=unsafe-best-match
                        
# 2026-11-01 17:28:08  0.00B 设置环境变量 UV_HTTP_TIMEOUT
ENV UV_HTTP_TIMEOUT=500
                        
# 2026-11-01 17:28:08  56.96MB 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c python3 -m pip install uv # buildkit
                        
# 2026-11-01 17:28:06  0.00B 设置环境变量 CRIU_CUDA_PLUGIN_DIR
ENV CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu
                        
# 2026-11-01 17:28:06  23.56MB 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c if [ "${TARGETPLATFORM}" = "linux/amd64" ]         && [ "${CUDA_VERSION%%.*}" -ge 13 ]; then         MAX_JOBS="${max_jobs}" /usr/local/bin/install_snapshot_runtime.sh;     else         echo "Initialized snapshots are not installed for ${TARGETPLATFORM} with CUDA ${CUDA_VERSION}";     fi # buildkit
                        
# 2026-11-01 17:27:50  3.94KB 复制新文件或目录到容器中
COPY tools/install_snapshot_runtime.sh /usr/local/bin/install_snapshot_runtime.sh # buildkit
                        
# 2026-11-01 17:27:50  0.00B 设置环境变量 LD_LIBRARY_PATH
ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib64
                        
# 2026-11-01 17:27:50  0.00B 设置环境变量 PATH
ENV PATH=/usr/local/cuda/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
                        
# 2026-11-01 17:27:50  0.00B 设置环境变量 CUDA_HOME
ENV CUDA_HOME=/usr/local/cuda
                        
# 2026-11-01 17:27:50  2.95GB 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c CUDA_VERSION_DASH=$(echo $CUDA_VERSION | cut -d. -f1,2 | tr '.' '-') &&     apt-get update -y &&     apt-get install -y --no-install-recommends --allow-change-held-packages         cuda-nvcc-${CUDA_VERSION_DASH}         cuda-cudart-${CUDA_VERSION_DASH}         cuda-nvrtc-${CUDA_VERSION_DASH}         cuda-nvrtc-dev-${CUDA_VERSION_DASH}         cuda-cuobjdump-${CUDA_VERSION_DASH}         libcurand-dev-${CUDA_VERSION_DASH}         libcublas-dev-${CUDA_VERSION_DASH}         libnuma-dev         numactl &&     rm -rf /var/lib/apt/lists/* # buildkit
                        
# 2026-11-01 17:26:57  648.86MB 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c apt-get update -y     && apt-get install -y --no-install-recommends         ca-certificates         software-properties-common         curl         sudo         ffmpeg         libsm6         libxext6         libgl1     && if [ ! -z ${DEADSNAKES_MIRROR_URL} ] ; then         if [ ! -z "${DEADSNAKES_GPGKEY_URL}" ] ; then             mkdir -p -m 0755 /etc/apt/keyrings ;             curl -L ${DEADSNAKES_GPGKEY_URL} | gpg --dearmor > /etc/apt/keyrings/deadsnakes.gpg ;             sudo chmod 644 /etc/apt/keyrings/deadsnakes.gpg ;             echo "deb [signed-by=/etc/apt/keyrings/deadsnakes.gpg] ${DEADSNAKES_MIRROR_URL} $(lsb_release -cs) main" > /etc/apt/sources.list.d/deadsnakes.list ;         fi ;     else         for i in 1 2 3; do             add-apt-repository -y ppa:deadsnakes/ppa && break ||             { echo "Attempt $i failed, retrying in 5s..."; sleep 5; };         done ;     fi     && apt-get update -y     && apt-get install -y --no-install-recommends         python${PYTHON_VERSION}         python${PYTHON_VERSION}-dev         python${PYTHON_VERSION}-venv         libibverbs-dev     && rm -rf /var/lib/apt/lists/*     && update-alternatives --install /usr/bin/python3 python3 /usr/bin/python${PYTHON_VERSION} 1     && update-alternatives --set python3 /usr/bin/python${PYTHON_VERSION}     && ln -sf /usr/bin/python${PYTHON_VERSION}-config /usr/bin/python3-config     && rm -f /usr/lib/python${PYTHON_VERSION}/EXTERNALLY-MANAGED     && curl -sS ${GET_PIP_URL} | python${PYTHON_VERSION}     && python3 --version && python3 -m pip --version # buildkit
                        
# 2026-11-01 17:25:46  136.00B 执行命令并创建新的镜像层
RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/amd64 max_jobs=16 CUDA_VERSION=13.0.2 PYTHON_VERSION=3.12 NCCL_VERSION=2.31.2 DEADSNAKES_MIRROR_URL= DEADSNAKES_GPGKEY_URL= GET_PIP_URL=https://bootstrap.pypa.io/get-pip.py /bin/sh -c PYTHON_VERSION_STR=$(echo ${PYTHON_VERSION} | sed 's/\.//g') &&     echo "export PYTHON_VERSION_STR=${PYTHON_VERSION_STR}" >> /etc/environment # buildkit
                        
# 2026-11-01 17:25:43  0.00B 设置工作目录为/vllm-workspace
WORKDIR /vllm-workspace
                        
# 2026-11-01 17:25:43  0.00B 设置环境变量 DEBIAN_FRONTEND
ENV DEBIAN_FRONTEND=noninteractive
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG GET_PIP_URL
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG DEADSNAKES_GPGKEY_URL
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG DEADSNAKES_MIRROR_URL
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG NCCL_VERSION
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG PYTHON_VERSION
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG CUDA_VERSION
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG max_jobs=2
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG TARGETPLATFORM
                        
# 2026-11-01 17:25:43  0.00B 定义构建参数
ARG INSTALL_RUBIN_PRERELEASE
                        
# 2025-10-11 00:48:26  0.00B 设置环境变量 NVIDIA_DRIVER_CAPABILITIES
ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility
                        
# 2025-10-11 00:48:26  0.00B 设置环境变量 NVIDIA_VISIBLE_DEVICES
ENV NVIDIA_VISIBLE_DEVICES=all
                        
# 2025-10-11 00:48:26  17.29KB 复制新文件或目录到容器中
COPY NGC-DL-CONTAINER-LICENSE / # buildkit
                        
# 2025-10-11 00:48:26  0.00B 设置环境变量 LD_LIBRARY_PATH
ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib64
                        
# 2025-10-11 00:48:26  0.00B 设置环境变量 PATH
ENV PATH=/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
                        
# 2025-10-11 00:48:26  22.00B 执行命令并创建新的镜像层
RUN |1 TARGETARCH=amd64 /bin/sh -c echo "/usr/local/cuda/lib64" >> /etc/ld.so.conf.d/nvidia.conf # buildkit
                        
# 2025-10-11 00:48:26  322.86MB 执行命令并创建新的镜像层
RUN |1 TARGETARCH=amd64 /bin/sh -c apt-get update && apt-get install -y --no-install-recommends     cuda-cudart-13-0=${NV_CUDA_CUDART_VERSION}     cuda-compat-13-0     && rm -rf /var/lib/apt/lists/* # buildkit
                        
# 2025-10-11 00:48:09  0.00B 设置环境变量 CUDA_VERSION
ENV CUDA_VERSION=13.0.2
                        
# 2025-10-11 00:48:09  10.26MB 执行命令并创建新的镜像层
RUN |1 TARGETARCH=amd64 /bin/sh -c apt-get update && apt-get install -y --no-install-recommends     gnupg2 curl ca-certificates &&     curl -fsSL https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/${NVARCH}/3bf863cc.pub | apt-key add - &&     echo "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2404/${NVARCH} /" > /etc/apt/sources.list.d/cuda.list &&     apt-get purge --autoremove -y curl     && rm -rf /var/lib/apt/lists/* # buildkit
                        
# 2025-10-11 00:48:09  0.00B 添加元数据标签
LABEL maintainer=NVIDIA CORPORATION <cudatools@nvidia.com>
                        
# 2025-10-11 00:48:09  0.00B 定义构建参数
ARG TARGETARCH
                        
# 2025-10-11 00:48:09  0.00B 设置环境变量 NV_CUDA_CUDART_VERSION
ENV NV_CUDA_CUDART_VERSION=13.0.96-1
                        
# 2025-10-11 00:48:09  0.00B 设置环境变量 NVIDIA_REQUIRE_CUDA brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand brand
ENV NVIDIA_REQUIRE_CUDA=cuda>=13.0 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<536 brand=nvidia,driver>=535,driver<536 brand=quadro,driver>=535,driver<536 brand=quadrortx,driver>=535,driver<536 brand=nvidiartx,driver>=535,driver<536 brand=vapps,driver>=535,driver<536 brand=vpc,driver>=535,driver<536 brand=vcs,driver>=535,driver<536 brand=vws,driver>=535,driver<536 brand=cloudgaming,driver>=535,driver<536 brand=unknown,driver>=550,driver<551 brand=grid,driver>=550,driver<551 brand=tesla,driver>=550,driver<551 brand=nvidia,driver>=550,driver<551 brand=quadro,driver>=550,driver<551 brand=quadrortx,driver>=550,driver<551 brand=nvidiartx,driver>=550,driver<551 brand=vapps,driver>=550,driver<551 brand=vpc,driver>=550,driver<551 brand=vcs,driver>=550,driver<551 brand=vws,driver>=550,driver<551 brand=cloudgaming,driver>=550,driver<551 brand=unknown,driver>=565,driver<566 brand=grid,driver>=565,driver<566 brand=tesla,driver>=565,driver<566 brand=nvidia,driver>=565,driver<566 brand=quadro,driver>=565,driver<566 brand=quadrortx,driver>=565,driver<566 brand=nvidiartx,driver>=565,driver<566 brand=vapps,driver>=565,driver<566 brand=vpc,driver>=565,driver<566 brand=vcs,driver>=565,driver<566 brand=vws,driver>=565,driver<566 brand=cloudgaming,driver>=565,driver<566 brand=unknown,driver>=570,driver<571 brand=grid,driver>=570,driver<571 brand=tesla,driver>=570,driver<571 brand=nvidia,driver>=570,driver<571 brand=quadro,driver>=570,driver<571 brand=quadrortx,driver>=570,driver<571 brand=nvidiartx,driver>=570,driver<571 brand=vapps,driver>=570,driver<571 brand=vpc,driver>=570,driver<571 brand=vcs,driver>=570,driver<571 brand=vws,driver>=570,driver<571 brand=cloudgaming,driver>=570,driver<571 brand=unknown,driver>=575,driver<576 brand=grid,driver>=575,driver<576 brand=tesla,driver>=575,driver<576 brand=nvidia,driver>=575,driver<576 brand=quadro,driver>=575,driver<576 brand=quadrortx,driver>=575,driver<576 brand=nvidiartx,driver>=575,driver<576 brand=vapps,driver>=575,driver<576 brand=vpc,driver>=575,driver<576 brand=vcs,driver>=575,driver<576 brand=vws,driver>=575,driver<576 brand=cloudgaming,driver>=575,driver<576
                        
# 2025-10-11 00:48:09  0.00B 设置环境变量 NVARCH
ENV NVARCH=x86_64
                        
# 2025-10-01 21:01:37  0.00B 
/bin/sh -c #(nop)  CMD ["/bin/bash"]
                        
# 2025-10-01 21:01:37  78.12MB 
/bin/sh -c #(nop) ADD file:249778a1782b02a1c2bcf9f292f5778d81442a53c3de1958d712f10baf7e0b60 in / 
                        
# 2025-10-01 21:01:35  0.00B 
/bin/sh -c #(nop)  LABEL org.opencontainers.image.version=24.04
                        
# 2025-10-01 21:01:35  0.00B 
/bin/sh -c #(nop)  LABEL org.opencontainers.image.ref.name=ubuntu
                        
# 2025-10-01 21:01:35  0.00B 
/bin/sh -c #(nop)  ARG LAUNCHPAD_BUILD_ARCH
                        
# 2025-10-01 21:01:35  0.00B 
/bin/sh -c #(nop)  ARG RELEASE
                        
                    

镜像信息

{
    "Id": "sha256:c139d638fa4bff3513333b742d7b5baf6267a04454f3e20e2310185e28241387",
    "RepoTags": [
        "gpustack/runner:cuda13.0-vllm0.33.0",
        "swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner:cuda13.0-vllm0.33.0"
    ],
    "RepoDigests": [
        "gpustack/runner@sha256:bbebaaa0e42bce48bfa7ed302f08984793afd417b65ccc72c5e8f3d4589ea610",
        "swr.cn-north-4.myhuaweicloud.com/ddn-k8s/docker.io/gpustack/runner@sha256:fb197394d3dea8f1a07605358866a34a406c3e8f342fc19b5d15cec8e32db31e"
    ],
    "Parent": "",
    "Comment": "buildkit.dockerfile.v0",
    "Created": "2026-11-04T11:47:07.278969633Z",
    "Container": "",
    "ContainerConfig": null,
    "DockerVersion": "",
    "Author": "",
    "Config": {
        "Hostname": "",
        "Domainname": "",
        "User": "",
        "AttachStdin": false,
        "AttachStdout": false,
        "AttachStderr": false,
        "Tty": false,
        "OpenStdin": false,
        "StdinOnce": false,
        "Env": [
            "PATH=/usr/local/cuda/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin",
            "NVARCH=x86_64",
            "NVIDIA_REQUIRE_CUDA=cuda\u003e=13.0 brand=unknown,driver\u003e=535,driver\u003c536 brand=grid,driver\u003e=535,driver\u003c536 brand=tesla,driver\u003e=535,driver\u003c536 brand=nvidia,driver\u003e=535,driver\u003c536 brand=quadro,driver\u003e=535,driver\u003c536 brand=quadrortx,driver\u003e=535,driver\u003c536 brand=nvidiartx,driver\u003e=535,driver\u003c536 brand=vapps,driver\u003e=535,driver\u003c536 brand=vpc,driver\u003e=535,driver\u003c536 brand=vcs,driver\u003e=535,driver\u003c536 brand=vws,driver\u003e=535,driver\u003c536 brand=cloudgaming,driver\u003e=535,driver\u003c536 brand=unknown,driver\u003e=550,driver\u003c551 brand=grid,driver\u003e=550,driver\u003c551 brand=tesla,driver\u003e=550,driver\u003c551 brand=nvidia,driver\u003e=550,driver\u003c551 brand=quadro,driver\u003e=550,driver\u003c551 brand=quadrortx,driver\u003e=550,driver\u003c551 brand=nvidiartx,driver\u003e=550,driver\u003c551 brand=vapps,driver\u003e=550,driver\u003c551 brand=vpc,driver\u003e=550,driver\u003c551 brand=vcs,driver\u003e=550,driver\u003c551 brand=vws,driver\u003e=550,driver\u003c551 brand=cloudgaming,driver\u003e=550,driver\u003c551 brand=unknown,driver\u003e=565,driver\u003c566 brand=grid,driver\u003e=565,driver\u003c566 brand=tesla,driver\u003e=565,driver\u003c566 brand=nvidia,driver\u003e=565,driver\u003c566 brand=quadro,driver\u003e=565,driver\u003c566 brand=quadrortx,driver\u003e=565,driver\u003c566 brand=nvidiartx,driver\u003e=565,driver\u003c566 brand=vapps,driver\u003e=565,driver\u003c566 brand=vpc,driver\u003e=565,driver\u003c566 brand=vcs,driver\u003e=565,driver\u003c566 brand=vws,driver\u003e=565,driver\u003c566 brand=cloudgaming,driver\u003e=565,driver\u003c566 brand=unknown,driver\u003e=570,driver\u003c571 brand=grid,driver\u003e=570,driver\u003c571 brand=tesla,driver\u003e=570,driver\u003c571 brand=nvidia,driver\u003e=570,driver\u003c571 brand=quadro,driver\u003e=570,driver\u003c571 brand=quadrortx,driver\u003e=570,driver\u003c571 brand=nvidiartx,driver\u003e=570,driver\u003c571 brand=vapps,driver\u003e=570,driver\u003c571 brand=vpc,driver\u003e=570,driver\u003c571 brand=vcs,driver\u003e=570,driver\u003c571 brand=vws,driver\u003e=570,driver\u003c571 brand=cloudgaming,driver\u003e=570,driver\u003c571 brand=unknown,driver\u003e=575,driver\u003c576 brand=grid,driver\u003e=575,driver\u003c576 brand=tesla,driver\u003e=575,driver\u003c576 brand=nvidia,driver\u003e=575,driver\u003c576 brand=quadro,driver\u003e=575,driver\u003c576 brand=quadrortx,driver\u003e=575,driver\u003c576 brand=nvidiartx,driver\u003e=575,driver\u003c576 brand=vapps,driver\u003e=575,driver\u003c576 brand=vpc,driver\u003e=575,driver\u003c576 brand=vcs,driver\u003e=575,driver\u003c576 brand=vws,driver\u003e=575,driver\u003c576 brand=cloudgaming,driver\u003e=575,driver\u003c576",
            "NV_CUDA_CUDART_VERSION=13.0.96-1",
            "CUDA_VERSION=13.0.3",
            "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/lib64",
            "NVIDIA_VISIBLE_DEVICES=all",
            "NVIDIA_DRIVER_CAPABILITIES=compute,utility",
            "DEBIAN_FRONTEND=noninteractive",
            "CUDA_HOME=/usr/local/cuda",
            "CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu",
            "UV_HTTP_TIMEOUT=500",
            "UV_INDEX_STRATEGY=unsafe-best-match",
            "UV_LINK_MODE=copy",
            "UV_PYTHON_INSTALL_DIR=/opt/uv/python",
            "UV_CACHE_DIR=/opt/uv/cache",
            "UV_COMPILE_BYTECODE=1",
            "UV_OVERRIDE=/etc/uv-overrides.txt",
            "VLLM_ENABLE_CUDA_COMPATIBILITY=0",
            "TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 8.9 9.0 10.0 12.0",
            "VLLM_USAGE_SOURCE=production-docker-image",
            "VLLM_BUILD_COMMIT=4ba787ec9959e566d3b587a796dff07673324da5",
            "VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b",
            "VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/9673",
            "VLLM_IMAGE_TAG=vllm/vllm-openai:v0.33.0",
            "LANG=en_US.UTF-8",
            "LANGUAGE=en_US:en",
            "LC_ALL=en_US.UTF-8",
            "PYTHON_VERSION=3.12",
            "PIP_NO_CACHE_DIR=1",
            "PIP_DISABLE_PIP_VERSION_CHECK=1",
            "PIP_ROOT_USER_ACTION=ignore",
            "CUDA_ARCHS=",
            "UV_SYSTEM_PYTHON=1",
            "UV_PRERELEASE=allow",
            "UV_SKIP_WHEEL_FILENAME_CHECK=1",
            "UV_NO_CACHE=1",
            "VLLM_VERSION=0.33.0",
            "VLLM_TORCH_VERSION=2.14.0",
            "VLLM_TORCH_CUDA_VERSION=13.0.3",
            "VLLM_LMCACHE_VERSION=0.5.6",
            "RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES=1",
            "SAFETENSORS_FAST_GPU=1",
            "RUNAI_STREAMER_MEMORY_LIMIT=0",
            "RUNAI_STREAMER_LOG_TO_STDERR=1",
            "RUNAI_STREAMER_LOG_LEVEL=INFO"
        ],
        "Cmd": null,
        "Image": "",
        "Volumes": null,
        "WorkingDir": "/",
        "Entrypoint": [
            "tini",
            "--"
        ],
        "OnBuild": null,
        "Labels": {
            "ai.vllm.build.commit": "4ba787ec9959e566d3b587a796dff07673324da5",
            "ai.vllm.build.pipeline": "019d130e-464e-4ff7-b84b-492992c0c06b",
            "ai.vllm.build.url": "https://buildkite.com/vllm/release-v2/builds/9673",
            "ai.vllm.image.tag": "vllm/vllm-openai:v0.33.0",
            "maintainer": "NVIDIA CORPORATION \u003ccudatools@nvidia.com\u003e",
            "org.opencontainers.image.ref.name": "ubuntu",
            "org.opencontainers.image.revision": "4ba787ec9959e566d3b587a796dff07673324da5",
            "org.opencontainers.image.source": "https://github.com/vllm-project/vllm",
            "org.opencontainers.image.url": "https://buildkite.com/vllm/release-v2/builds/9673",
            "org.opencontainers.image.version": "vllm/vllm-openai:v0.33.0"
        },
        "Shell": [
            "/bin/bash",
            "-eo",
            "pipefail",
            "-c"
        ]
    },
    "Architecture": "amd64",
    "Os": "linux",
    "Size": 27466383909,
    "GraphDriver": {
        "Data": {
            "LowerDir": "/var/lib/docker/overlay2/290b8daa509067a4208633af5149b7fffa7220b6caf99eb6943a4e2dd9cf42b3/diff:/var/lib/docker/overlay2/17c7689df85ab6aecd42dc80cfec8e7116fd4cc3e516ecc1b899626c9ba4658e/diff:/var/lib/docker/overlay2/78670c7b1c49564aa3f3401cc61fe90b4be7e5364b09f78e9a62dda1d6363cc4/diff:/var/lib/docker/overlay2/3b296a30684a096567c6802ced12864017b1921b136a57a68587b1068790591d/diff:/var/lib/docker/overlay2/4d64c48b61aa95691fa46a8ccf812872c9c186cc3d3b9c0cd4aaa5a912e373f0/diff:/var/lib/docker/overlay2/9f396d57f93880ac86ad13243de229c27d5180b1af8d64eb0ebb961c562e5ce2/diff:/var/lib/docker/overlay2/f290c467c33371c67ba936fb28deed3af0e26a815751dda83af549d437eaf724/diff:/var/lib/docker/overlay2/ce8b87ba52f04dd9e96d96f6f404c96f5fa72c48617409498d5336ff07ee6cd2/diff:/var/lib/docker/overlay2/632408097eaa4599e859958887463cbb067b251c193ec1e6c2621a4e2e2ff466/diff:/var/lib/docker/overlay2/2ba93a51aadc241baf67f646bee52a3790f5ad19ce7b28f261d1bd9b0cdcfd70/diff:/var/lib/docker/overlay2/db4c9cc75d9e814c70e65d1b646f0ff6ad11f5a2db8696220ac7a71a1bb57f9c/diff:/var/lib/docker/overlay2/1fa00dc3872436259554a88b0cc48a18f34841e6e73781f56c9575a5879cc8b0/diff:/var/lib/docker/overlay2/76959e9eff80dd8d7fdcbe44f97d539431554e8c173004b151ba78c4da3d2634/diff:/var/lib/docker/overlay2/de3d4b18dafa9c92724e6e41efd37664e65704833e46a42753c978e6fde6bdec/diff:/var/lib/docker/overlay2/1a8b77eae402e7581dab5f964c0e19e32b19006f0c9f1a2492408617597bf352/diff:/var/lib/docker/overlay2/14c4672a7f1818a68cd74d700a5f39b3da0f36dab9f9ca43088f7212897fb393/diff:/var/lib/docker/overlay2/dc791ea98d8ba483afc5a471fb5594e8b551ae12127517d7be46ff6c0b9b9838/diff:/var/lib/docker/overlay2/abb14301709c8669e9d7f3cc3d91fed934f3e194f2e13cdb4a600ac5b7688361/diff:/var/lib/docker/overlay2/cbe492a5dd21f288c5c90bae5780eaa61b6be523d4e734331691d5ddf5c1f68e/diff:/var/lib/docker/overlay2/d30fcf007586518800912ff17432c00e73730adc1aa24240e210c2984cfd1024/diff:/var/lib/docker/overlay2/e592c18edd13a5a5af79f1981dec8b91844763802f103b6340e6f2c761e0ae9c/diff:/var/lib/docker/overlay2/90ac8ff9f1db9548d2a8f87e1ad5d22c37cec080b6c237ef97f68c296fd17e84/diff:/var/lib/docker/overlay2/e47823586dbc59e9ed473efa8ccf48ecdcafb4395f0cb75102f8f50b9b223c03/diff:/var/lib/docker/overlay2/4aa86e46d1abf652e899d1dbb6ef2b1416588750382ae707414493e532da566c/diff:/var/lib/docker/overlay2/dde292910f14967a9b866b4d893bdb80e5715f006621d5dd615b3763b918455e/diff:/var/lib/docker/overlay2/798dd96c45642ab436f0d6350d59e6e27db87f66c4a504b07851ccd4605728e1/diff:/var/lib/docker/overlay2/a7e3205895334c1e686c9f4c909feb5768a94d1e26b7a0109969e944e933b607/diff:/var/lib/docker/overlay2/06a748a7ab788808c0d56f0f8fd2b6bdf638e0862a28dbbdc604e42bd724baeb/diff:/var/lib/docker/overlay2/ac61765739bc08ab31c82b5f3784653f586946f26db0d5e2be6eb318517b2148/diff:/var/lib/docker/overlay2/49ea9fc8d95d5225df644408f3538989bfd50308e794c78e13b3554c3d69ec9a/diff:/var/lib/docker/overlay2/4403a3f3f6e4142350df0ac1211e3128d75ac4e6bdaf9c7d37ab03e7d2b867bc/diff:/var/lib/docker/overlay2/253259e99f0b95a0d55dd3780f8ef0b08d02ff554576a7a937a0ac5d664248d4/diff:/var/lib/docker/overlay2/421b6de28e83bce697867b7e3393300e2cb4ec2ddee87491b3f278bf75ce8d41/diff:/var/lib/docker/overlay2/71ace43eaef8d3c1aa8e33330816610113382f5acdc3ae2373eb12317ed542b5/diff:/var/lib/docker/overlay2/420c602b2e98de070a2ddd27b98225e8ffc64d53b064a871fa3ccb73a4670876/diff:/var/lib/docker/overlay2/211c9497cbdfa35585cee3fb9b792f461e1913642f005091946dca09b8372bf8/diff:/var/lib/docker/overlay2/0b313e9585a747113dfb4578a9e6f1d14405ba5005ec5c885cc66f96d09d0a03/diff:/var/lib/docker/overlay2/83d67a79d6ce1aedfd781ada437d732068e3cd6762d93924c61920a56755472b/diff:/var/lib/docker/overlay2/679a33b30574c278f1748bad35d28294dc055b6a57c03cc047f50dd1f1029434/diff:/var/lib/docker/overlay2/8b999fc47ad6bd8eb3e63f1a9dcacf4c5cab93483410d4a12c1ae922d585f4a4/diff:/var/lib/docker/overlay2/2e2c2e9d68f62b0ffecb127af21be35cfdef76552a348060ca81ef03933d539d/diff:/var/lib/docker/overlay2/12f49c87da7b93f71ae3536bf357a02b12fde67a9fd0360950fbf0f60a12b062/diff:/var/lib/docker/overlay2/384f71387736bca4f2369777e565b9e9730a1837c4681b0ca2075f8208289e75/diff:/var/lib/docker/overlay2/bdfb668a0d0c798c12082d904620faec81e89174fe7040be11f618d7382ff95d/diff:/var/lib/docker/overlay2/444f4467e99a8822e0fb96b5926ab0e7f109bd83c55f47fe145c5bd815f02630/diff:/var/lib/docker/overlay2/d0b7b7b6c1fa613d44be3e89eaa7491cf03f2b993d329f4c008a59e32e73c340/diff:/var/lib/docker/overlay2/282bd9b031c557e5f38273cdc04a803ad01abc34edc9476ade7bddfce909512b/diff:/var/lib/docker/overlay2/cbc1a60ffcaed05ffb08a6d139d89875c18b069bb09de5d40938ed4193eeec4f/diff:/var/lib/docker/overlay2/3575796dd6e095d155e3024e1c5dab72ed6d80f679f03b363afbd50a712804dc/diff:/var/lib/docker/overlay2/b4f6fe9b2bfef766c4c58e129013cc05dabf80e46713b5708713bccd0e6cab77/diff:/var/lib/docker/overlay2/fb3ec169c24085cc551bfef52b4f2388a39d1478a85bef53f00eaf454f17ffae/diff",
            "MergedDir": "/var/lib/docker/overlay2/f315dbbbb62e7dea10643a4109a897593dd1c9e2818e6fbf8fb2f3486d4410c9/merged",
            "UpperDir": "/var/lib/docker/overlay2/f315dbbbb62e7dea10643a4109a897593dd1c9e2818e6fbf8fb2f3486d4410c9/diff",
            "WorkDir": "/var/lib/docker/overlay2/f315dbbbb62e7dea10643a4109a897593dd1c9e2818e6fbf8fb2f3486d4410c9/work"
        },
        "Name": "overlay2"
    },
    "RootFS": {
        "Type": "layers",
        "Layers": [
            "sha256:073ec47a8c22dcaa4d6e5758799ccefe2f9bde943685830b1bf6fd2395f5eabc",
            "sha256:18241330a2d180465d0b9868e4978456293d37da715de2f313ba08862da2aab1",
            "sha256:b5c4de646bbad7dd50f917376023047d99ff9a1e806943dd86799e933b6a6294",
            "sha256:dab2da33fe40a74d9faf939b829b701aa6cfc6b4502ff68fcf716fd2973697ac",
            "sha256:b7ec8fd7c485bbedbe64cbc5a7a810ceabb9156e95c83ba60596ddb9b4955d48",
            "sha256:7f889496e83e6174ce4c8655ac5dcf9c7209b0fb8e7023029fc25288032342b3",
            "sha256:86a4568520b614af374c6102e4a313d29207948d289324ed6f3609e73989f93a",
            "sha256:ca7c4b03f85e5eea6dba822458a3d5dffbcdd5039f1c64141dd64eb644c5f8f9",
            "sha256:ca3152014295eb09c1a519cd9a21a65d063d69dd6ddefdff82f825cbee75f23c",
            "sha256:9a4a16b220a5276b4b5eb0c652705c560fff872e8677f860ea89a083dfa3398c",
            "sha256:d295c117000b806dc4463174d07407f4b9625c1e5321faf78be8165dc680363f",
            "sha256:ba99c2402a2dc9d00d24cb05d85a6630505554b6098f0d36b4888ffad44f6c91",
            "sha256:4b9b226b673b8ccdebc4b11de2521d6e228f0a415cd2734fd3bfcc9c37b5c8f5",
            "sha256:a39e2ab9fe995a481e983cdecadfcdd182c878abfc52b2b4e9a3c235542ba828",
            "sha256:8a5f88d01a8122f574bcf895b0c6ebedcb06ec3bcc36ccbec7ce630a11ea75e1",
            "sha256:84e8484199b7b59889a70a4e5e6b671ece03d13b7bdec007028a698529592096",
            "sha256:84e8484199b7b59889a70a4e5e6b671ece03d13b7bdec007028a698529592096",
            "sha256:a7c3edb6d6ccbdbba7c37c2ec921eb4c262accb446a972b46a1b72ce03b6a2af",
            "sha256:cccc2e484f062450b5bfa06b713f4f349c568de769375c19f04aa3511745861f",
            "sha256:5803ec82ef3211d487642a2a4a4285752684efbf086d8f601e149ee76088f716",
            "sha256:636f5b574001bb6425436a504f0906464d606fc867d4e43f33d6e28044999289",
            "sha256:021d881c8b594974492fd58f5925f4fce8486a0231995ee2af036e328d82d706",
            "sha256:7ff3313d5a174675e21d6aa5d49d74d0047d9b20799b03203a6f79117f75d100",
            "sha256:be4ba8c11e94cafe4c0356824e9ed66ece567bcf746b2af124568d73b3eaaf94",
            "sha256:4b6d3e5d80b7b712f154668a6afd7ca8b1e64c9246dc48ca1ed041690ccb5cd1",
            "sha256:e76ba55bb1d9528f3c99d0ef5927fcbac5d590b7eac1179aaaf884924b4d1462",
            "sha256:7f31a8ec76a3d6f1d26c7cf1a2d613f8e5c5b535b695713b88a03c05642fc0b4",
            "sha256:75353b7e9ddcbeecd37336aa9588d61bce10e1c2433287d9e78ec15946df597e",
            "sha256:5f70bf18a086007016e948b04aed3b82103a36bea41755b6cddfaf10ace3c6ef",
            "sha256:1bd74dded3dcd4fd374c4dd5ff4b4569a7ab8c56653bcd547a66eeff665ecacd",
            "sha256:6cf989e36800530a48e96108de3ce70903229d57b2fd259ca859b5189ebed9d8",
            "sha256:8f3e17df09d9f46126d0f96928d7f53ebc3b2c40dd6b3f53bc733e4d04363c37",
            "sha256:ec3ce27cc816203a153b5defc22a2385ecab609058b4d09dbde2743c9b4da348",
            "sha256:ec0524c906425eb3e6038d9286ddac521ed3db4b66f55f11a7f87a7888f8e7a8",
            "sha256:bef62a0ead347bc9fee8587a4aec09357fe175f5ee8150dcf022edeea92cfe5e",
            "sha256:45f3574df9c06fb2c85ec44b2162be0e03d1314f47f58f82f371b04c0824b8a1",
            "sha256:13a3e4fb7ad0828c750b545139c1ea63c8978a6c44b0953e1a92292fb164b0e6",
            "sha256:08c6187d9ef052bb30e139be8d8670ba76e1c63f976a19baf629b8957f88e69d",
            "sha256:5500e8f012d441826a3884e95af824b1f697649ff460107acb483a97dd419ad7",
            "sha256:50f916eb5129b1fcf29965ab02a2e7476355c1111e39fbf627edb6fc26676dae",
            "sha256:529dc06a3b51da4d7f4bf1cce157eaff5d14fbf8b3a75fb1b1699979ce9ecab4",
            "sha256:7dd33806f96f01d0ff7916cec70b4a8af5355e3c33c34f2ab16aadbaac2cd158",
            "sha256:0997b01e33aa316891704d96d8f47f7b4947b260630360419004de5bee67ceed",
            "sha256:efa9ffd3c0338af26ca320319a3965669bb4aea41b8c5ab755beec972f1649d1",
            "sha256:25a5a73ec39cff1ab39213843f0e48fce91ca997dcb70dc0cdc440c543dfb597",
            "sha256:50e3f282e10c3ef616cab789386e66f9b4c1f08362c1fd19d66c4fab1b35aa38",
            "sha256:452e7f4e4b3e9d7610741b224df5eccd830d7fe870e5d84c2937f8cd65285f2b",
            "sha256:2e826b0f6c5b267c80045dc599db5486e25763f12d627a7d01c96af0119ba35a",
            "sha256:5f70bf18a086007016e948b04aed3b82103a36bea41755b6cddfaf10ace3c6ef",
            "sha256:c9ad65e6cb01c50e9a29bc9ec929c780ac6a0d9a6a2c4fdea61a5da756f4938a",
            "sha256:a970e6c5842fda681faadd8431f6bd1d47bedad7d2838d1f6bc6faa6a1e9c71d",
            "sha256:5f70bf18a086007016e948b04aed3b82103a36bea41755b6cddfaf10ace3c6ef"
        ]
    },
    "Metadata": {
        "LastTagTime": "2026-11-19T00:42:48.536766554+08:00"
    }
}

更多版本

docker.io/gpustack/runner:cann8.2-910b-mindie2.1.rc2

linux/arm64 docker.io16.02GB2025-12-05 01:13
619

docker.io/gpustack/runner:cann8.2-910b-sglang0.5.2

linux/amd64 docker.io17.03GB2025-12-05 01:02
508

docker.io/gpustack/runner:cann8.2-910b-sglang0.5.2

linux/arm64 docker.io18.27GB2025-12-05 01:28
590

docker.io/gpustack/runner:cann8.2-910b-vllm0.11.0

linux/arm64 docker.io15.96GB2025-11-27 00:53
768

docker.io/gpustack/runner:cann9.0-910b-vllm0.20.2

linux/arm64 docker.io18.66GB2026-09-10 00:24
157

docker.io/gpustack/runner:cuda12.4-vllm0.11.0

linux/amd64 docker.io24.31GB2025-12-05 02:03
956

docker.io/gpustack/runner:cuda12.4-voxbox0.0.20

linux/amd64 docker.io17.15GB2025-12-05 04:14
398

docker.io/gpustack/runner:cuda12.8-sglang0.5.5

linux/amd64 docker.io32.96GB2025-12-05 02:43
862

docker.io/gpustack/runner:cuda12.8-voxbox0.0.21

linux/amd64 docker.io20.57GB2026-03-05 00:39
453

docker.io/gpustack/runner:cuda12.9-vllm0.11.2

linux/amd64 docker.io33.62GB2025-12-03 00:42
890

docker.io/gpustack/runner:cuda12.9-vllm0.19.1

linux/amd64 docker.io23.47GB2026-05-12 02:14
344

docker.io/gpustack/runner:cuda12.9-vllmv0.20.1

linux/amd64 docker.io27.44GB2026-05-09 02:28
407

docker.io/gpustack/runner:cuda13.0-vllm0.22.1

linux/amd64 docker.io27.16GB2026-06-26 00:41
357

docker.io/gpustack/runner:cuda13.0-vllm0.25.1

linux/amd64 docker.io20.37GB2026-08-09 02:16
271

docker.io/gpustack/runner:cuda13.0-vllm0.27.1

linux/amd64 docker.io23.38GB2026-09-11 01:34
176

docker.io/gpustack/runner:cuda13.0-vllm0.33.0

linux/amd64 docker.io27.47GB2026-11-19 00:44
9

docker.io/gpustack/runner:cuda13.0-vllmv0.20.1

linux/amd64 docker.io24.18GB2026-05-08 14:23
331

docker.io/gpustack/runner:rocm7.1-vllm0.17.1

linux/amd64 docker.io31.12GB2026-04-22 00:42
378
检测到您正在使用广告拦截插件,本站为公益站点,依赖广告维持运转 🙏 查看如何关闭 ×