The NVIDIA container image of TensorFlow, release 17.04, is available.
TensorFlow container image version 17.04 is based on TensorFlow 1.0.1.
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® 8.0.61
- NVIDIA CUDA® Deep Neural Network library™ (cuDNN) 6.0.20
- NVIDIA® Collective Communications Library ™ (NCCL) 1.6.1 (optimized for NVLink™ )
Key Features and Enhancements
This TensorFlow release includes the following key features and enhancements.
- 2x improvement on 8GPUs; 1.5X on 4 GPUs
- Updated NCCL integration with support for NVLink
- Multi-GPU CNN examples that demonstrates efficient training of CNNs using NCCL
- XLA (Accelerated Linear Algebra) support enabled, allowing users to offload operations to TensorFlow experimental XLA back-end
- Ubuntu 16.04 with March 2017 updates
Known Issues
There are no known issues in this release.