PyTorch Release 18.12

The NVIDIA container image for PyTorch, release 18.12, 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 10.0.130 including CUDA^® Basic Linear Algebra Subroutines library™ (cuBLAS) 10.0.130
• NVIDIA CUDA^® Deep Neural Network library™ (cuDNN) 7.4.1
• NCCL 2.3.7 (optimized for NVLink™ )
• APEx
• OpenMPI 3.1.2
• Caffe2
• TensorRT 5.0.2
• DALI 0.5.0 Beta
• Tensor Core Optimized Examples:
  • ResNet50 v1.5
  • GNMT v2

Driver Requirements

Release 18.12 is based on CUDA 10, which requires NVIDIA Driver release 410.xx. However, if you are running on Tesla (Tesla V100, Tesla P4, Tesla P40, or Tesla P100), you may use NVIDIA driver release 384. For more information, see CUDA Compatibility and Upgrades.

GPU Requirements

Release 18.12 supports CUDA compute capability 6.0 and higher. This corresponds to GPUs in the Pascal, Volta, and Turing families. Specifically, for a list of GPUs that this compute capability corresponds to, see CUDA GPUs. For additional support details, see Deep Learning Frameworks Support Matrix.

Key Features and Enhancements

This PyTorch release includes the following key features and enhancements.

• PyTorch container image version 18.12 is based on PyTorch v0.4.1+ with up-to-date features from the PyTorch v1.0 preview (main branch up to PR 12303). PyTorch 0.4.1+ is released and included with this container.
• Performance improvement for PyTorch’s native batch normalization.
• Mixed precision SoftMax enabling FP16 inputs, FP32 computations and FP32 outputs.
• Latest version of DALI 0.5.0 Beta.
• Ubuntu 16.04 with November 2018 updates

Tensor Core Examples

• An implementation of ResNet50. The ResNet50 v1.5 model is a modified version of the original ResNet50 v1 model.
• An implementation of GNMT v2. The GNMT v2 model is similar to the one discussed in the Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation paper.

Known Issues

Persistent batch normalization kernels have been disabled due to a known bug during validation. Batch normalization provides correct results and work as expected from users, however, this may cause up to 10% regression in time to solution performance on networks using batch normalization.