TensorFlow Release 17.11
The NVIDIA container image of TensorFlow, release 17.11, is available.
TensorFlow container image version 17.11 is based on TensorFlow 1.3.0.
Contents of TensorFlow
This container image contains the complete source of the version of NVIDIA TensorFlow in
/opt/tensorflow. It is pre-built and installed into the
/usr/local/[bin,lib] directories in the container image.
To achieve optimum TensorFlow performance, for image based training, the container includes a sample script that demonstrates efficient training of convolutional neural networks (CNNs). The sample script may need to be modified to fit your application. The container also includes the following:
- Ubuntu 16.04
- NVIDIA CUDA 9.0.176 including CUDA® Basic Linear Algebra Subroutines library™ (cuBLAS) 9.0.234
- NVIDIA CUDA® Deep Neural Network library™ (cuDNN) 7.0.4
- NVIDIA® Collective Communications Library ™ (NCCL) 2.1.2 (optimized for NVLink™ )
Release 17.11 is based on CUDA 9, which requires NVIDIA Driver release 384.xx.
Key Features and Enhancements
This TensorFlow release includes the following key features and enhancements.
- Added support for float16 data type and Tensor Core math in batched matrix multiply operations.
- Added example script
nvidia-examples/cnn/nvcnn_hvd.py, which demonstrates use of the Horovod library for multi-node training.
Dockerfile.horovoddemonstrating how to build a Docker container with the Horovod library and MPI support.
- Added OpenSeq2Seq example demonstrating sequence-to-sequence model training in
- Latest version of CUDA
- Latest version of cuDNN
- Latest version of NCCL
- Ubuntu 16.04 with October 2017 updates
There are no known issues in this release.