NVIDIA
NVIDIA
Dynamo vLLM Runtime
Container
NVIDIA
NVIDIA
Dynamo vLLM Runtime

The Dynamo vLLM runtime image is a containerized build of Dynamo + vLLM which serves as the base runtime environment for vLLM based inference with Dynamo's distributed inference framework.

LayerLabelCreated
1a11ed7172bee1e6dfc3fb91b801a1524f26a3a187691986cde2f29ed465e0ddCONFIG
Entrypoint /opt/nvidia/nvidia_entrypoint.sh; WorkingDir /workspace
05/06/2026 6:45 PM UTC
900c838028cf7529357e96033b06296ec8bf89d91ce1d6b8b5ca11b6828867ccENTRYPOINT
/opt/nvidia/nvidia_entrypoint.sh
05/06/2026 6:45 PM UTC
4fbd68388aff0e15838484e87448cd3f3dfd2a4ab1b214c0e5981a52ceb23350ENV
VLLM_USE_FLASHINFER_SAMPLER=1
05/06/2026 6:45 PM UTC
0df1a38618123c4f18e659a6d24f91d3b4af5ae542c8af3fb5d95f23b8deebbdENV
DYNAMO_COMMIT_SHA=64ca94114f019dc7a87cd4cd08cdd82943763c83
05/06/2026 6:45 PM UTC
da2a8de5d52bc28843c5dc5b331ef7ba3c0a56fcaf9c5cc4f83accd37311414bARG
DYNAMO_COMMIT_SHA=64ca94114f019dc7a87cd4cd08cdd82943763c83
05/06/2026 6:45 PM UTC
ba1d8256a78d29f787e7e66cb158c45a9c42cbf1f522d744f350dd4d56c34754USER
dynamo
05/06/2026 6:45 PM UTC
101c0d6e904cc489d861933c2ef88ed77766106ca8f117968008b041505b4caaRUN
DEVICE=cuda PYTHON_VERSION=3.12 SITE_PACKAGES=/opt/dynamo/venv/lib/python3.12/site-packages ENABLE_KVBM=true ENABLE_GPU_MEMORY_SERVICE=true ENABLE_MODELEXPRESS_P2P=false MODELEXPRESS_REF=76fc5d7f06c37121ee8789a29fac6f9b08c4743a /bin/bash -l -o pipefail -c cd /usr/local/lib &&
  if [ -f libaws-c-common.so.1.0.0 ] &&
  [ ! -L libaws-c-common.so.1 ]; then rm -f libaws-c-common.so.1 libaws-c-common.so &&
  ln -s libaws-c-common.so.1.0.0 libaws-c-common.so.1 &&
  ln -s libaws-c-common.so.1 libaws-c-common.so; fi &&
  if [ -f libaws-c-s3.so.1.0.0 ] &&
  [ ! -L libaws-c-s3.so.0unstable ]; then rm -f libaws-c-s3.so.0unstable libaws-c-s3.so &&
  ln -s libaws-c-s3.so.1.0.0 libaws-c-s3.so.0unstable &&
  ln -s libaws-c-s3.so.0unstable libaws-c-s3.so; fi &&
  if [ -f libs2n.so.1.0.0 ] &&
  [ ! -L libs2n.so.1 ]; then rm -f libs2n.so.1 libs2n.so &&
  ln -s libs2n.so.1.0.0 libs2n.so.1 &&
  ln -s libs2n.so.1 libs2n.so; fi &&
  for lib in libcrypto libssl; do versioned=$(ls -1 ${lib}.so.1.1.* 2>/dev/null | head -1); if [ -n "$versioned" ] &&
  [ ! -L "${lib}.so.1.1" ]; then rm -f "${lib}.so.1.1" &&
  ln -s "$(basename "$versioned")" "${lib}.so.1.1"; fi; done &&
  ldconfig
05/06/2026 6:45 PM UTC
a321b9a4e5842c7d3a703862fa73c3bd5008e08271acd50836ae5f08899a1fddCOPY
--chown=dynamo: /usr/lib64/libssl.so.1.1* /usr/local/lib/
05/06/2026 6:45 PM UTC
a73336b4b082978523753784e69b4aca4292b8315df16c69bf7ee6e05ab5dcabCOPY
--chown=dynamo: /usr/lib64/libcrypto.so.1.1* /usr/local/lib/
05/06/2026 6:45 PM UTC
4e7296fbe20c7ac9ac4ef1f113a43a4870338beec6b236c1fab95f64b3276c0aCOPY
--chown=dynamo: /usr/local/lib64/libs2n* /usr/local/lib/
05/06/2026 6:45 PM UTC
...

NVIDIA uses cookies to improve your experience on our web site. We and our third-party partners also use cookies and other tools to collect and record information you provide as well as information about your interactions with our websites for performance improvement, analytics, and to assist in marketing efforts. By clicking "Accept All", you consent to our use of cookies and other tools as described in our Cookie Policy. You can manage your cookie settings by clicking on "Manage Settings." By continuing to use this site or by clicking one of the buttons below, you agree to our Terms of Service (which contains important waivers). Please see our Privacy Policy for more information on our privacy practices.