DALI Release 0.13.0 Beta

The DALI 0.13.0 is a beta release. Hence, for all the features, the functionality and performance will likely be limited.

Using DALI 0.13.0 Beta

The DALI 0.13.0 can be used with the 19.08 NVIDIA GPU Cloud (NGC) optimized container for MXNet, PyTorch, and TensorFlow. Also, the 19.08 container will be shipped with DALI 0.13.0.

To upgrade to DALI 0.13.0 beta from an older version of DALI, follow the installation instructions in the DALI Quick Start Guide.

Refer to the DALI Developer Guide for usage details.

Note: The internal DALI C++ API used for operators implementation, and the C++ API that enables using DALI as a library from native code, are not yet officially supported. Hence these APIs may change in the next release without advance notice.

Key Features and Enhancements

This DALI release includes the following key features and enhancements.

  • Added new:
    • Fast coco reader
    • TorchPythonFunction operator
    • Sink operators
  • Reworked how the reader pick samples from the shuffling buffer.
  • Used DALI_extra for test data.
  • Added checks to see if Python API is not mixed between simple, scheduled and iterator APIs.
  • Added support for reading video files with labels using file_list argument.

Fixed Issues

This DALI release includes the following fixes.

  • Restored support of use_batched_decode argument in nvJPEGDecoder operator (only for legacy nvJPEGDecoder implementation).
  • Fixed FP16 support in DALI TensorFlow plugin.
  • Fixed Python operator with side effects.
  • Fixed a race condition in async pipeline executor.
  • Disabled video_reader_op test when NVDEC is disabled.
  • Fixed sampling of chroma in the VideoReader operation.
  • Fix detection pipeline example.

Breaking Changes

  • Reader sampling from shuffling buffer was adjusted. Now samples are not mixed between epochs.

Deprecated Features

  • Deprecated NormalizePermute in favor of CropMirrorNormalize
  • Multiple Input Sets handling was removed from backend and is only Python level syntactic sugar.

Known Issues

  • The new video reader operator requires NVIDIA VIDEO CODEC SDK support in the platform. Prior to 19.01, the NVIDIA GPU Cloud (NGC) optimized containers lack this functionality in the default configuration. To enable the functionality, run the container with the "video" capability enabled, as shown below: 
    -e "NVIDIA_DRIVER_CAPABILITIES=compute,utility,video"
  • The video loader operator requires that the key frames occur at a minimum every 10 to 15 frames of the video stream. If the key frames occur at a lesser frequency, then the returned frames may be out of sync.
  • DALI TensorFlow plugin may be not compatible with TensorFlow 1.14.0 release. The DALI TensorFlow plugin requires that the gcc compiler that matches the one used to build TensorFlow (gcc 4.8.4 or gcc 4.8.5, depending on the particular version) be present on the system.