DALI Release 0.5 Beta

This DALI 0.5 release is a beta release.

Key Features and Enhancements

This DALI release includes the following key features and enhancements.
  • Fixed rare hangs during DALI pipeline destruction.

  • Added ResNet50 example with TensorFlow.

  • Added more verbose error messages to HostDecoder and nvJpegDecoder.

  • Added pipelined execution completion callback setter.

  • Added support for Tiff images.

Breaking API Changes

  • Random seed type changed from INT to INT64, therefore, serialized pipelines from versions prior to 0.5 are not compatible with the current DALI version.

Using DALI 0.5 Beta

The 18.11 NVIDIA GPU Cloud (NGC) optimized container for MXNet, PyTorch, and TensorFlow, includes an older version of DALI. To upgrade to DALI 0.5 beta, follow the installation instructions in the DALI Quick Start Guide.

Refer to the DALI Developer Guide for usage details.

Known Issues

  • The DALI integrated ResNet-50 samples in the 18.11 NGC TensorFlow and PyTorch containers may result in lower than expected performance results. We are working to address the issue in the next release.

  • This is a beta release. All features are expected to be available, however, some aspects of functionality and performance will likely be limited compared to a non-beta release.

  • AWS performance in the TensorFlow example is lower than expected without DALI. The reason is a too small number of threads is assigned to DALI. This will be resolved in a future release.