ContainersNVIDIADeepSeek-R1NVIDIALlama-3.1-Nemotron-70B-InstructNVIDIAPyTorch
CollectionsNVIDIAOmniverse Kit (FB)NVIDIADeepStream SDKNVIDIAOmniverse Kit App Streaming
ModelsNVIDIAStyleGAN3 pretrained modelsNVIDIAPeopleNetNVIDIATrafficCamNet
ResourcesNVIDIAKit SDK - Windows (PB25h1)NVIDIAKit SDK - Linux (PB25h1)NVIDIARiva Skills Quick Start
Helm ChartsNVIDIAGPU OperatorNVIDIANVIDIA NIM Operator
NVIDIA
NVIDIA
Holoscan Container
Container
NVIDIA
NVIDIA
Holoscan Container

The Holoscan container includes the Holoscan libraries, GXF extensions, headers, example source code, and sample datasets. It is the recommended way to run the Holoscan examples or build your own applications.

NVIDIA AI Enterprise Supported
LayerLabelCreated
727b56b0e805f42ef1bdcf71d605ad8f3e589e22dd26fa0254c4241eeea72d88CONFIG
Entrypoint /opt/nvidia/nvidia_entrypoint.sh; WorkingDir /opt/nvidia/holoscan
11/25/2025 12:39 AM UTC
8f837c840771006b5c59380f33241797832e1af56f74d900749f90e749e43210COPY
docker/release/entrypoint.d/ /opt/nvidia/entrypoint.d/
11/25/2025 12:39 AM UTC
3dddb69bb79bacc575e426ffa135e36be47f1386b04cab1afaeedf2c3aa9edcaENV
NVIDIA_PRODUCT_NAME=NVIDIA Holoscan SDK
11/25/2025 12:39 AM UTC
2af673a37f435fa03a5b9e3a9e33cdf7ec114d702f13c0bdb383b36d8115d298ENV
HOLOSCAN_SDK_VERSION=3.9.0
11/25/2025 12:39 AM UTC
b82509a82b4a9b8651e164cce4de3b5970723003aa8e6446abbd188952088a2bARG
HOLOSCAN_SDK_VERSION=3.9.0
11/25/2025 12:39 AM UTC
15b2fe5fa93531cc884ef6f96e3e30d943c8d46de4c2f7c51af8daeb89a40a87RUN
ENABLE_APT_CACHING=true DEBIAN_FRONTEND=noninteractive TARGETARCH=amd64 GPU_TYPE=dgpu DOCA_VERSION=3.1.0 INSTALL_PATH=/opt/nvidia/holoscan SAMPLE_DATA_PATH=/opt/nvidia/holoscan/data NCCL_VERSION=2.27 CUDA_MAJOR=13 CCCL_VERSION=3.0.1 HOST_INSTALL_DIR=install-cu13-x86_64 HOST_SDK_INSTALL_DIR=/workspace/holoscan-sdk ONNX_RUNTIME_VERSION=1.22.1 CUDSS_VERSION=0.6.0.5 mkdir /opt/nvidia/legal &&
  mv /opt/nvidia/entrypoint.d/13-tensorrt-copyright.txt /opt/nvidia/entrypoint.d/30-container-license.txt /opt/nvidia/legal/ ||
  true
11/25/2025 12:39 AM UTC
5f3351a7142320146bdb551d11aa94e807d809a5c6316b37437095654438b7ddRUN
ENABLE_APT_CACHING=true DEBIAN_FRONTEND=noninteractive TARGETARCH=amd64 GPU_TYPE=dgpu DOCA_VERSION=3.1.0 INSTALL_PATH=/opt/nvidia/holoscan SAMPLE_DATA_PATH=/opt/nvidia/holoscan/data NCCL_VERSION=2.27 CUDA_MAJOR=13 CCCL_VERSION=3.0.1 HOST_INSTALL_DIR=install-cu13-x86_64 HOST_SDK_INSTALL_DIR=/workspace/holoscan-sdk ONNX_RUNTIME_VERSION=1.22.1 CUDSS_VERSION=0.6.0.5 rm -f /opt/nvidia/entrypoint.d/40-tensorrt-samples-info.txt /opt/nvidia/entrypoint.d/29-tensorrt-url.txt /opt/nvidia/entrypoint.d/12-banner.sh ||
  true
11/25/2025 12:39 AM UTC
98ed40f244cda24086b0e73e53234a261df2ca65f215489183bba01b6e260d69RUN
ENABLE_APT_CACHING=true DEBIAN_FRONTEND=noninteractive TARGETARCH=amd64 GPU_TYPE=dgpu DOCA_VERSION=3.1.0 INSTALL_PATH=/opt/nvidia/holoscan SAMPLE_DATA_PATH=/opt/nvidia/holoscan/data NCCL_VERSION=2.27 CUDA_MAJOR=13 CCCL_VERSION=3.0.1 HOST_INSTALL_DIR=install-cu13-x86_64 HOST_SDK_INSTALL_DIR=/workspace/holoscan-sdk ONNX_RUNTIME_VERSION=1.22.1 CUDSS_VERSION=0.6.0.5 if find ${LIBTORCH} -name libtorch.so | xargs -I {} ldd {} | grep -q "libcudss.so"; then curl -S -# -o cudss.tar.xz "https://developer.download.nvidia.com/compute/cudss/redist/libcudss/linux-$(uname -m)/libcudss-linux-$(uname -m)-${CUDSS_VERSION}_cuda${CUDA_MAJOR}-archive.tar.xz" &&
  tar xJf cudss.tar.xz --strip-components 1 --no-same-owner --no-same-permissions &&
  rm -f cudss.tar.xz &&
  echo "/opt/nvidia/cudss/lib" >> /etc/ld.so.conf.d/cudss.conf &&
  ldconfig &&
  ldconfig -p | grep -q cudss; fi
11/25/2025 12:39 AM UTC
1d3b14dd30be10e82da4a8618b79f8a58e5b2ca9aaf33c8ed9f055bf63e03656WORKDIR
/opt/nvidia/cudss
11/25/2025 12:39 AM UTC
fa5613d88ff3f8e252d6e1fc9d68c8880113984ea90a734f4f3f5e21a6a92373ARG
CUDA_MAJOR=13
11/25/2025 12:39 AM UTC
...