Sign inSign up

clustermax/cmax-inference:0.1.0

Index digest

sha256:2d530ff7e65b350ab88b6a2d9460d631bef1b1c16eb0cc5521605dea56c3c28a

OS/ARCH

linux/arm64

Compressed size

11.96 GB

Last pushed

3 months by clustermax

Type

Image

Manifest digest

sha256:e28e1fc472bd8223c4282ce61bc39e7640fbe37d564ce52b6afb831cf4296952

Image Layers

Image Layers

1ARG RELEASE0 B
2ARG LAUNCHPAD_BUILD_ARCH0 B
3LABEL org.opencontainers.image.ref.name=ubuntu0 B
4LABEL org.opencontainers.image.version=24.040 B
5ADD file ... in / 28.55 MB
6CMD ["/bin/bash"]0 B
7ENV NVARCH=sbsa0 B
8ENV NVIDIA_REQUIRE_CUDA=cuda>=12.9 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<5360 B
9ENV NV_CUDA_CUDART_VERSION=12.9.79-10 B
10ARG TARGETARCH0 B
11LABEL maintainer=NVIDIA CORPORATION <[email protected]>0 B
12RUN |1 TARGETARCH=arm64 /bin/sh -c4.32 MB
13ENV CUDA_VERSION=12.9.10 B
14RUN |1 TARGETARCH=arm64 /bin/sh -c93.06 MB
15RUN |1 TARGETARCH=arm64 /bin/sh -c182 B
16ENV PATH=/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin0 B
17ENV LD_LIBRARY_PATH=/usr/local/cuda/lib640 B
18COPY NGC-DL-CONTAINER-LICENSE / # buildkit6.72 KB
19ENV NVIDIA_VISIBLE_DEVICES=all0 B
20ENV NVIDIA_DRIVER_CAPABILITIES=compute,utility0 B
21ARG CUDA_VERSION0 B
22ARG PYTHON_VERSION0 B
23ARG DEADSNAKES_MIRROR_URL0 B
24ARG DEADSNAKES_GPGKEY_URL0 B
25ARG GET_PIP_URL0 B
26ENV DEBIAN_FRONTEND=noninteractive0 B
27WORKDIR /vllm-workspace102 B
28RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=230 B
29RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=215.92 MB
30RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=2.09 GB
31RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=45.75 MB
32ENV UV_HTTP_TIMEOUT=5000 B
33ENV UV_INDEX_STRATEGY=unsafe-best-match0 B
34ENV UV_LINK_MODE=copy0 B
35ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python0 B
36ENV UV_CACHE_DIR=/opt/uv/cache0 B
37RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=150 B
38RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=3.62 KB
39COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit2.03 KB
40RUN |5 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=2.03 KB
41ENV VLLM_ENABLE_CUDA_COMPATIBILITY=00 B
42ARG PYTORCH_CUDA_INDEX_BASE_URL0 B
43COPY requirements/common.txt /tmp/common.txt # buildkit1.59 KB
44COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit697 B
45RUN |6 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=4.65 GB
46ARG FLASHINFER_VERSION=0.6.11.post20 B
47RUN |7 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=1.86 GB
48ARG GDRCOPY_CUDA_VERSION=12.80 B
49ARG GDRCOPY_OS_VERSION=Ubuntu22_040 B
50ARG TARGETPLATFORM0 B
51COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit897 B
52RUN |10 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=518.2 KB
53ARG BITSANDBYTES_VERSION_X86=0.46.10 B
54ARG BITSANDBYTES_VERSION_ARM64=0.42.00 B
55ARG TIMM_VERSION=>=1.0.170 B
56ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.70 B
57RUN |14 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=360.63 MB
58ARG PIP_INDEX_URL UV_INDEX_URL0 B
59ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL0 B
60ARG PYTORCH_CUDA_INDEX_BASE_URL0 B
61ARG PIP_KEYRING_PROVIDER UV_KEYRING_PROVIDER0 B
62ARG PYTORCH_NIGHTLY0 B
63COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit200 B
64RUN |22 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=688.19 MB
65RUN |22 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=32 B
66ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda/lib640 B
67RUN |22 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=15.21 MB
68RUN |22 CUDA_VERSION=12.9.1 PYTHON_VERSION=3.12 DEADSNAKES_MIRROR_URL=348.34 MB
69ENV LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib640 B
70COPY examples examples # buildkit254.38 KB
71COPY benchmarks benchmarks # buildkit239.49 KB
72COPY ./vllm/collect_env.py . # buildkit9.62 KB
73ARG TARGETPLATFORM0 B
74ARG INSTALL_KV_CONNECTORS=false0 B
75ARG CUDA_VERSION0 B
76ARG VLLM_BUILD_COMMIT0 B
77ARG VLLM_BUILD_PIPELINE0 B
78ARG VLLM_BUILD_URL0 B
79ARG VLLM_IMAGE_TAG0 B
80ARG PIP_INDEX_URL UV_INDEX_URL0 B
81ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL0 B
82ENV UV_HTTP_TIMEOUT=5000 B
83ARG torch_cuda_arch_list=7.5 8.0 8.6 8.90 B
84ENV TORCH_CUDA_ARCH_LIST=8.0 8.7 8.9 9.00 B
85RUN |12 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true CUDA_VERSION=12.9.11.12 GB
86ARG MOONCAKE_WHEEL_AARCH640 B
87ARG MOONCAKE_WHEEL_X86_640 B
88RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true CUDA_VERSION=12.9.1425.34 MB
89ENV VLLM_USAGE_SOURCE=production-docker-image0 B
90ENV VLLM_BUILD_COMMIT=0decac0d96c42b49572498019f0a0e3600f50398 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/2333 VLLM_IMAGE_TAG=vllm/vllm-openai:v0.22.1-cu129-ubuntu24040 B
91LABEL org.opencontainers.image.source=https://github.com/vllm-project/vllm org.opencontainers.image.revision=0decac0d96c42b49572498019f0a0e3600f50398 org.opencontainers.image.version=vllm/vllm-openai:v0.22.1-cu129-ubuntu2404 org.opencontainers.image.url=https://buildkite.com/vllm/release-v2/builds/23330 B
92ENTRYPOINT ["vllm" "serve"]0 B
93LABEL ai.clustermax.image=cmax-inference ai.clustermax.base=vllm/vllm-openai:v0.22.1-aarch64-cu129-ubuntu2404 ai.clustermax.base.digest=sha256:4e5d0d839a292df9ce438d73bea96b8ec61a4fcd96f31005f3aa9ddf85f63f3f ai.clustermax.vllm.version=0.22.10 B
94RUN /bin/sh -c pip install78.02 MB

Command

 ARG RELEASE