TensorFlow Release 18.12

The NVIDIA container image of TensorFlow, release 18.12, is available.

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 as a system Python module.

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
  Note:

  Container image 18.12-py2 contains Python 2.7; 18.12-py3 contains Python 3.5.

• 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™ )
• Horovod 0.15.1
• OpenMPI 3.1.2
• TensorBoard 1.12.0
• MLNX_OFED 3.4
• OpenSeq2Seq v18.12 at commit 59c70e7
• TensorRT 5.0.2
• DALI 0.5.0 Beta

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 TensorFlow release includes the following key features and enhancements.

• TensorFlow container image version 18.12 is based on TensorFlow 1.12.0.
• Latest version of DALI 0.5.0 Beta.
• OpenSeq2Seq’s custom CTC decoder is now pre-built in the container.
• The tensorflow.contrib.nccl module has been moved into core as tensorflow.python.ops.nccl_ops. User scripts may need to be updated accordingly. No changes are required for Horovod users. For an example of using Horovod, refer to the nvidia-examples/cnn/ directory.
• Inference image classification examples have been removed from the container and are now available at: GitHub: TensorFlow/TensorRT Integration.
• Ubuntu 16.04 with November 2018 updates

Accelerating Inference In TensorFlow With TensorRT (TF-TRT)

For step-by-step instructions on how to use TF-TRT, see Accelerating Inference In TensorFlow With TensorRT User Guide.

Deprecated Features

• The image-classification examples were moved from /opt/tensorflow/nvidia-examples/inference/image-classification/scripts to https://github.com/tensorflow/tensorrt/tree/master/tftrt/examples/image-classification.

• The check_accuracy.py script, used to check whether the accuracy generated by the example matches with the expectation, was removed from the example. Refer to the published accuracy numbers to verify whether your generated accuracy numbers match with the expectation.

Announcements

Support for accelerating TensorFlow with TensorRT 3.x will be removed in a future release (likely TensorFlow 1.13). The generated plan files are not portable across platforms or TensorRT versions. Plans are specific to the exact GPU model they were built on (in addition to the platforms and the TensorRT version) and must be retargeted to the specific GPU in case you want to run them on a different GPU. Therefore, models that were accelerated using TensorRT 3.x will no longer run. If you have a production model that was accelerated with TensorRT 3.x, you will need to convert your model with TensorRT 4.x or later again.

For more information, see the Note in Serializing A Model In C++ or Serializing A Model In Python.

Known Issues

• OpenSeq2Seq is only supported in the Python 3 container.
• Horovod and XLA cannot be used together due to a known issue in upstream TensorFlow. We expect this to be resolved in an upcoming release.