PyTorch Release 18.07
The NVIDIA container image of PyTorch, release 18.07, is available.
Contents of PyTorch
This container image contains the complete source of the version of PyTorch in
/opt/pytorch. It is pre-built and installed in the
pytorch-py3.6 Conda™ environment in the container image.
The container also includes the following:
- Ubuntu 16.04 including Python 3.6 environment
- NVIDIA CUDA 9.0.176 (see Errata section and 2.1) including CUDA® Basic Linear Algebra Subroutines library™ (cuBLAS) 9.0.425
- NVIDIA CUDA® Deep Neural Network library™ (cuDNN) 7.1.4
- NCCL 2.2.13 (optimized for NVLink™ )
- Caffe2 0.8.1
- DALI 0.1 Beta
Release 18.07 is based on CUDA 9, which requires NVIDIA Driver release 384.xx.
Key Features and Enhancements
- PyTorch container image version 18.07 is based on PyTorch 0.4.0 upstream main branch post commit cca2476.
- Clip grads can be used on a single tensor directly.
- The precision of MSELoss with half inputs has been improved.
- PyTorch's JIT (still in Alpha) now supports FP16 inputs and outputs, comparisons, the exp operator, and ReLU gates.
- Added support for DALI 0.1 Beta.
- Latest version of CUDA® Basic Linear Algebra Subroutines library™ (cuBLAS) 9.0.425.
- Ubuntu 16.04 with June 2018 updates
When importing Caffe2 after importing Torch, there is an issue which causes GPU support for Caffe2 to be disabled. For users affected by this bug, it is recommended to either use the PyTorch 18.06 or 18.08 container.