Torch Release 16.12

The NVIDIA container image of Torch, release 16.12, is available.

Torch container image version 16.12 is based on Torch7.

Contents of Torch

This container image contains the complete source of the version of NVIDIA Torch in /opt/torch. It is pre-built and installed into the /usr/local/[bin,share,lib] directories in the container image.

The NVIDIA® Collective Communications Library ™ (NCCL) (NCCL) library and Torch bindings for NCCL are installed in this container, and models using DataParallelTable can easily leverage this library for fast parallel training.

The container also includes the following:

Key Features and Enhancements

This Torch release includes the following key features and enhancements.
  • Supports FP32 and FP16 storage and FP32 arithmetic
    • Pseudo-FP16 support in cuDNN layers
  • Optimized multi-GPU training
    • Seamless NCCL integration (opt-in)
    • Better control over workspace memory usage via support the new cudnnFindEx() routine; enable in model script by adding cudnn.useFindEx = true
  • Supports recurrent neural networks (RNNs)
    • Supports cuDNN RNN layers
    • Tuned RNN performance
  • Lua frontend
  • Ubuntu 14.04 with November 2016 updates

Known Issues

There are no known issues in this release.