ContainersNVIDIADeepSeek-R1NVIDIALlama-3.1-Nemotron-70B-InstructNVIDIAPyTorch
CollectionsNVIDIAOmniverse Kit (FB)NVIDIADeepStream SDKNVIDIAOmniverse Kit App Streaming
ModelsNVIDIAStyleGAN3 pretrained modelsNVIDIAPeopleNetNVIDIATrafficCamNet
ResourcesNVIDIARiva Skills Quick Start
Helm ChartsNVIDIAGPU OperatorNVIDIANVIDIA NIM Operator
NVIDIA
NVIDIA
PyTorch
Container
NVIDIA
NVIDIA
PyTorch

PyTorch is a GPU accelerated tensor computational framework. Functionality can be extended with common Python libraries such as NumPy and SciPy. Automatic differentiation is done with a tape-based system at the functional and neural network layer levels.

NVIDIA AI Enterprise Supported
LayerLabelCreated
49806280968efa02e30f0b19f61df24193483d9a12c253894a2c8b73d9bef87fCONFIG
Entrypoint /opt/nvidia/nvidia_entrypoint.sh; WorkingDir /workspace; ExposedPorts 6006/tcp, 8888/tcp
04/15/2025 8:00 PM UTC
2394e55c4a05bccd6b8c98e8aa5237fb866cbde725b7501c9054fbb5e2d8e63bARG
NVIDIA_BUILD_REF=8d28f2ff75a3f186220e187654c5890a49f8d87a
04/15/2025 8:00 PM UTC
31b5bf64a6f776ff273407eab01081127ba5eba4ae4a3a263cee73d9d99f6a85LABEL
com.nvidia.build.id=159049541
04/15/2025 8:00 PM UTC
7aa0d04760bbb01b7dfb579d1192ef73c871ad307dd16a76c9ca5fcd67f07023ENV
NVIDIA_BUILD_ID=159049541
04/15/2025 8:00 PM UTC
8ce593f5c36c61bf24eb891ed4b8dedc667c1008dee0c71a5cacb3c5eb1d9333ARG
NVIDIA_BUILD_ID=159049541
04/15/2025 8:00 PM UTC
3ebfdd73ab786b27ddf88cdca589ceb81941eb2b7bd29f7a36b0b3ecc041f303COPY
entrypoint.d/ /opt/nvidia/entrypoint.d/
04/15/2025 8:00 PM UTC
dc90dbf758f53f9c34fca433501ed9f978be518d7416e8adb76d0ca0e4bdf8c5RUN
NVIDIA_PYTORCH_VERSION=25.04 PYTORCH_BUILD_VERSION=2.7.0a0+79aa174 NVFUSER_BUILD_VERSION=5111d3b TARGETARCH=amd64 PYVER=3.12 PYVER_MAJMIN=312 L4T=0 DALI_EXTRA_INDEX_URL=http://sqrl/nvdl/datasets/dali/pip-dali ln -sf ${_CUDA_COMPAT_PATH}/lib.real ${_CUDA_COMPAT_PATH}/lib &&
  echo ${_CUDA_COMPAT_PATH}/lib > /etc/ld.so.conf.d/00-cuda-compat.conf &&
  ldconfig &&
  rm -f ${_CUDA_COMPAT_PATH}/lib
04/15/2025 8:00 PM UTC
bbb3701ec316675d4c592abe9770002a32d941b592823cb1b3fee782d32aef9cENV
TORCH_NCCL_USE_COMM_NONBLOCKING=0
04/15/2025 8:00 PM UTC
e31bef6a25f4a37539e802fb93e845bb27c912351e8f946bd30c530fef2b9f30ENV
CUDA_MODULE_LOADING=LAZY
04/15/2025 8:00 PM UTC
f026e7c6ce3ae79bd00c987929a9e8d484cc3528817b877f54b555b681fee043RUN
NVIDIA_PYTORCH_VERSION=25.04 PYTORCH_BUILD_VERSION=2.7.0a0+79aa174 NVFUSER_BUILD_VERSION=5111d3b TARGETARCH=amd64 PYVER=3.12 PYVER_MAJMIN=312 L4T=0 DALI_EXTRA_INDEX_URL=http://sqrl/nvdl/datasets/dali/pip-dali if [ "${L4T}" = "1" ]; then echo "Not installing Transformer Engine in iGPU container until Version variable is set"; else /nvidia/build-scripts/installCAL.sh &&
  /nvidia/build-scripts/installCUBLASMP.sh &&
  /nvidia/build-scripts/installNVSHMEM.sh &&
  git clone -b release_v${TRANSFORMER_ENGINE_VERSION} --single-branch --recursive https://github.com/NVIDIA/TransformerEngine.git &&
  env NVTE_CUDA_ARCHS="70;80;89;90;100;120" NVTE_BUILD_THREADS_PER_JOB=8 pip install --no-cache-dir --no-build-isolation ./TransformerEngine &&
  rm -rf TransformerEngine; fi
04/15/2025 8:00 PM UTC
...