Sign inSign up

vllm/vime:latest

Multi-platform
Index digest

sha256:fa7ce3d660fa9e096f2d88b3947cb8502f5946ef3a33dab4a609c5d10f5c333f

OS/ARCH

Compressed size

14.75 GB

Last pushed

1 day by vllm

Type

Image

Manifest digest

sha256:fa7fac3d59cce98b29b4c83f668cfe54a1111eda39e3837fab05a78dfe5cfe6a

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>=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=13.0.96-10 B
10ARG TARGETARCH0 B
11LABEL maintainer=NVIDIA CORPORATION <[email protected]>0 B
12RUN |1 TARGETARCH=arm64 /bin/sh -c4.32 MB
13ENV CUDA_VERSION=13.0.20 B
14RUN |1 TARGETARCH=arm64 /bin/sh -c95.11 MB
15RUN |1 TARGETARCH=arm64 /bin/sh -c186 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_PRERELEASE0 B
22ARG TARGETPLATFORM0 B
23ARG max_jobs=20 B
24ARG CUDA_VERSION0 B
25ARG PYTHON_VERSION0 B
26ARG NCCL_VERSION0 B
27ARG DEADSNAKES_MIRROR_URL0 B
28ARG DEADSNAKES_GPGKEY_URL0 B
29ARG GET_PIP_URL0 B
30ENV DEBIAN_FRONTEND=noninteractive0 B
31WORKDIR /vllm-workspace104 B
32RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16230 B
33RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16216.18 MB
34RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=161.49 GB
35ENV CUDA_HOME=/usr/local/cuda0 B
36ENV 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
37ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib640 B
38COPY tools/install_snapshot_runtime.sh /usr/local/bin/install_snapshot_runtime.sh # buildkit1.62 KB
39RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=1632 B
40ENV CRIU_CUDA_PLUGIN_DIR=/usr/local/lib/criu0 B
41RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=1638.67 MB
42ENV UV_HTTP_TIMEOUT=5000 B
43ENV UV_INDEX_STRATEGY=unsafe-best-match0 B
44ENV UV_LINK_MODE=copy0 B
45ENV UV_PYTHON_INSTALL_DIR=/opt/uv/python0 B
46ENV UV_CACHE_DIR=/opt/uv/cache0 B
47ENV UV_COMPILE_BYTECODE=10 B
48RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16150 B
49RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16164 B
50ENV UV_OVERRIDE=/etc/uv-overrides.txt0 B
51RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=163.62 KB
52COPY docker/entrypoints/vllm-nonroot-entrypoint.sh /usr/local/bin/vllm-nonroot-entrypoint.sh # buildkit2.03 KB
53RUN |9 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=162.03 KB
54ENV VLLM_ENABLE_CUDA_COMPATIBILITY=00 B
55ARG PYTORCH_CUDA_INDEX_BASE_URL0 B
56COPY requirements/common.txt /tmp/common.txt # buildkit1.67 KB
57COPY requirements/cuda.txt /tmp/requirements-cuda.txt # buildkit866 B
58RUN |10 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=165 GB
59ARG FLASHINFER_VERSION=0.7.0.post10 B
60RUN |11 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=161.42 GB
61ARG GDRCOPY_CUDA_VERSION=12.80 B
62ARG GDRCOPY_OS_VERSION=Ubuntu24_040 B
63ARG TARGETPLATFORM0 B
64COPY tools/install_gdrcopy.sh /tmp/install_gdrcopy.sh # buildkit892 B
65RUN |14 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16517.81 KB
66ARG TIMM_VERSION=>=1.0.170 B
67ARG RUNAI_MODEL_STREAMER_VERSION=>=0.15.70 B
68RUN |16 INSTALL_RUBIN_PRERELEASE=false TARGETPLATFORM=linux/arm64 max_jobs=16288.65 MB
69ARG PIP_INDEX_URL UV_INDEX_URL0 B
70ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL0 B
71ARG PYTORCH_CUDA_INDEX_BASE_URL0 B
72ARG PIP_KEYRING_PROVIDER UV_KEYRING_PROVIDER0 B
73ARG PYTORCH_NIGHTLY0 B
74COPY /workspace/torch_lib_versions.txt torch_lib_versions.txt # buildkit198 B
75COPY /workspace/dist/wheel.sha256 /tmp/vllm-wheel.sha256 # buildkit241 B
76RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=531.95 MB
77RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=199 B
78RUN |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=32 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 |8 PIP_INDEX_URL= UV_INDEX_URL= PIP_EXTRA_INDEX_URL=13.71 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 # buildkit308.69 KB
84COPY benchmarks benchmarks # buildkit292.88 KB
85COPY ./vllm/collect_env.py . # buildkit9.61 KB
86ARG TARGETPLATFORM0 B
87ARG INSTALL_KV_CONNECTORS=false0 B
88ARG INSTALL_RUBIN_PRERELEASE0 B
89ARG TRITON_INSTALL_FROM_SOURCE_REPO0 B
90ARG CUDA_VERSION0 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_URL0 B
96ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL0 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=8.0 8.7 8.9 9.00 B
100RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false521.39 MB
101RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false92 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 # buildkit824 B
104RUN |14 TARGETPLATFORM=linux/arm64 INSTALL_KV_CONNECTORS=true INSTALL_RUBIN_PRERELEASE=false183 B
105ENV VLLM_USAGE_SOURCE=production-docker-image0 B
106ENV VLLM_BUILD_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 VLLM_BUILD_PIPELINE=019d130e-464e-4ff7-b84b-492992c0c06b VLLM_BUILD_URL=https://buildkite.com/vllm/release-v2/builds/7908 VLLM_IMAGE_TAG=vllm/vllm-openai:nightly-43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b50 B
107LABEL org.opencontainers.image.source=https://github.com/vllm-project/vllm org.opencontainers.image.revision=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 org.opencontainers.image.version=vllm/vllm-openai:nightly-43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 org.opencontainers.image.url=https://buildkite.com/vllm/release-v2/builds/79080 B
108ENTRYPOINT ["vllm" "serve"]0 B
109ARG PATCH_VERSION=latest0 B
110ARG VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b50 B
111ARG MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d0 B
112ARG DEEPGEMM_COMMIT=b38a77cd193cf38f670caae192310521d24343be0 B
113ARG DEEPEP_COMMIT=6845ffd9d59126ec0030c13e0e155935a61e5b5a0 B
114ARG DEEPEP_CUDA_ARCH_LIST=9.0;10.0;10.30 B
115ARG FLASH_QLA_COMMIT=821fd9d37ede18fdc2a4e707fefe3770bfc32e580 B
116ARG TRANSFORMER_ENGINE_VERSION=2.18.00 B
117ARG TRANSFORMER_ENGINE_COMMIT=27486e03cfc1fa41f6932dcecdc47c71c47eac3e0 B
118ARG TRANSFORMER_ENGINE_CUDA_ARCHS=90;100a;103a0 B
119ARG TMS_COMMIT=4d525cf378fdbfe7eb044909b8e66e112a5086240 B
120ARG TMS_CUDA_MAJOR=0 B
121ARG NCCL_VERSION=2.30.70 B
122ARG MAX_JOBS=160 B
123ENV MAX_JOBS=16 CMAKE_BUILD_PARALLEL_LEVEL=160 B
124ENV LD_LIBRARY_PATH=/usr/local/lib/python3.12/dist-packages/nvidia/nccl/lib:/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
125ARG ENABLE_CUDA_13=10 B
126ARG FA2_MAX_JOBS=320 B
127WORKDIR /root/32 B
128RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d133 B
129RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d1.72 GB
130RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d1 KB
131ENV PIP_BREAK_SYSTEM_PACKAGES=10 B
132RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d32 B
133RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d858.04 MB
134RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d378.82 MB
135RUN |16 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d4.02 MB
136ARG INSTALL_FLASHQLA=00 B
137RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d32 B
138RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d360 B
139RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d316.35 MB
140RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d49.88 MB
141RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d168.34 MB
142RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d904.82 KB
143RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d4.53 MB
144RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d11.61 MB
145RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d113.55 MB
146COPY requirements.txt /tmp/requirements.txt # buildkit452 B
147COPY /wheels/ /tmp/sunabako-wheels/ # buildkit810.7 KB
148RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d831.1 KB
149RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d767.44 MB
150RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d32 B
151RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d74.22 MB
152RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d156 B
153COPY docker/patch/latest/megatron*.patch /root/Megatron-LM/ # buildkit5.69 KB
154RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d152.15 MB
155COPY docker/patch/latest/vllm-pull_weights.patch /tmp/vllm-pull_weights.patch # buildkit4.39 KB
156COPY docker/patch/latest/vllm.patch /tmp/vllm.patch # buildkit2.59 KB
157COPY docker/patch/latest/vllm-score-centering.patch /tmp/vllm-score-centering.patch # buildkit4.42 KB
158COPY docker/patch/latest/vllm-pd-request-metrics.patch /tmp/vllm-pd-request-metrics.patch # buildkit5.39 KB
159COPY docker/patch/latest/vllm-inflight-queue-diagnostics.patch /tmp/vllm-inflight-queue-diagnostics.patch # buildkit2.19 KB
160COPY docker/patch/latest/vllm-retained-serving.patch /tmp/vllm-retained-serving.patch # buildkit2.26 KB
161COPY docker/patch/latest/vllm-generation-health.patch /tmp/vllm-generation-health.patch # buildkit1.25 KB
162COPY docker/patch/latest/vllm-tensor-dump.patch /tmp/vllm-tensor-dump.patch # buildkit27.25 KB
163COPY docker/patch/latest/vllm-fine-grained-tracing.patch /tmp/vllm-fine-grained-tracing.patch # buildkit5.85 KB
164COPY docker/patch/latest/vllm-nixl-peer-retirement.patch /tmp/vllm-nixl-peer-retirement.patch # buildkit3.22 KB
165COPY docker/patch/latest/vllm-mooncake-prom-metrics.patch /tmp/vllm-mooncake-prom-metrics.patch # buildkit2.4 KB
166COPY docker/patch/latest/vllm-host-cache-ratio.patch /tmp/vllm-host-cache-ratio.patch # buildkit718 B
167COPY docker/patch/latest/vllm-deepgemm-bf16.patch /tmp/vllm-deepgemm-bf16.patch # buildkit6.69 KB
168COPY docker/patch/latest/vllm-deepep-optional-shrink.patch /tmp/vllm-deepep-optional-shrink.patch # buildkit481 B
169RUN |17 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d529.04 KB
170COPY /build/rust/target/release/vllm-rs /usr/local/lib/python3.12/dist-packages/vllm/vllm-rs # buildkit22.46 MB
171ARG VIME_REPO=https://github.com/aoshen02/vime.git0 B
172ARG VIME_COMMIT=6addf1181c846158530bcd0d25e9338880ec799f0 B
173RUN |19 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d37.47 MB
174RUN |19 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d1.45 MB
175ENTRYPOINT []0 B
176RUN |19 PATCH_VERSION=latest VLLM_COMMIT=43b4aaea3e40e20ef53ce2bfecfa1b72060bf2b5 MEGATRON_COMMIT=4b4acac9a1d28ea6829c8d4f566d75698a21249d412.01 MB
177CMD ["/bin/bash"]0 B
178RUN /bin/sh -c rm -rf124 B
179COPY source/ /root/vime/ # buildkit5.71 MB
180COPY source.tar /root/vime/vime.tar # buildkit5.71 MB
181COPY engine/vllm/ /usr/local/lib/python3.12/dist-packages/vllm/ # buildkit72.95 KB
182COPY wheels/ /tmp/router-wheels/ # buildkit7.83 MB
183RUN /bin/sh -c uv pip16.37 MB
184RUN /bin/sh -c rm -rf128 B
185COPY source/ /root/vime/ # buildkit5.71 MB
186COPY source.tar /root/vime/vime.tar # buildkit5.71 MB
187COPY vllm-rs /usr/local/lib/python3.12/dist-packages/vllm/vllm-rs # buildkit22.49 MB
188RUN /bin/sh -c /usr/local/lib/python3.12/dist-packages/vllm/vllm-rs --version32 B
189RUN /bin/sh -c rm -rf124 B
190COPY source/ /root/vime/ # buildkit5.71 MB
191COPY source.tar /root/vime/vime.tar # buildkit5.71 MB
192RUN /bin/sh -c rm -rf124 B
193COPY source/ /root/vime/ # buildkit5.72 MB
194COPY source.tar /root/vime/vime.tar # buildkit5.72 MB
195COPY runtime/vllm/ /usr/local/lib/python3.12/dist-packages/vllm/ # buildkit572.92 KB
196COPY vllm-rs /usr/local/lib/python3.12/dist-packages/vllm/vllm-rs # buildkit22.49 MB
197RUN /bin/sh -c /usr/local/lib/python3.12/dist-packages/vllm/vllm-rs --version32 B
198RUN /bin/sh -c rm -rf126 B
199COPY source/ /root/vime/ # buildkit5.72 MB
200COPY source.tar /root/vime/vime.tar # buildkit5.72 MB

Command

 ARG RELEASE