PyTorch | NVIDIA NGC

NVIDIA

PyTorch

Container

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

Layer	Label		Created
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	LABEL	`com.nvidia.build.ref=25aae698093a80672c33aac448393478894f210e`	08/19/2024 7:48 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ARG	`NVIDIA_BUILD_REF=25aae698093a80672c33aac448393478894f210e`	08/19/2024 7:48 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	LABEL	`com.nvidia.build.id=107063152`	08/19/2024 7:48 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ENV	`NVIDIA_BUILD_ID=107063152`	08/19/2024 7:48 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ARG	`NVIDIA_BUILD_ID=107063152`	08/19/2024 7:48 AM UTC
sha256:e63ce922f0229bde5aea9f366c46883dcd23747e7d2c541f16665f199dbf98b8	COPY	`entrypoint.d/ /opt/nvidia/entrypoint.d/`	08/19/2024 7:48 AM UTC
sha256:87c7b4dbb6365d8608c348b35c20f9949546a6b42acdb47b491377f5d070dbe3	RUN	`NVIDIA_PYTORCH_VERSION=24.08 PYTORCH_BUILD_VERSION=2.5.0a0+872d972 NVFUSER_BUILD_VERSION=1d02b13 TARGETARCH=arm64 PYVER=3.10 L4T=1 TARGETARCH=arm64 PYVER=3.10 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 ${_CUDA_COMPAT_PATH}/lib`	08/19/2024 7:48 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ENV	`CUDA_MODULE_LOADING=LAZY`	08/19/2024 7:48 AM UTC
sha256:a3ed95caeb02ffe68cdd9fd84406680ae93d633cb16422d00e8a7c22955b46d4	ENV	`TORCH_CUDNN_V8_API_ENABLED=1`	08/19/2024 7:48 AM UTC
sha256:20d3c4ef0277522096f23258aa16249b49589c2d3810fbee02dec840b3b9c62d	RUN	`NVIDIA_PYTORCH_VERSION=24.08 PYTORCH_BUILD_VERSION=2.5.0a0+872d972 NVFUSER_BUILD_VERSION=1d02b13 TARGETARCH=arm64 PYVER=3.10 L4T=1 TARGETARCH=arm64 PYVER=3.10 if [ "${L4T}" = "1" ]; then echo "Not installing Transformer Engine in iGPU container because Version variable is set"; else pip install --no-cache-dir --no-build-isolation git+https://github.com/NVIDIA/TransformerEngine.git@release_v${TRANSFORMER_ENGINE_VERSION}; fi`	08/19/2024 7:48 AM UTC