DALI Release 0.12.0 Beta

The DALI 0.12.0 is a beta release. Hence, for all the features, the functionality and performance will likely be limited.

Using DALI 0.12.0 Beta

The DALI 0.12.0 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 0.12.0.

To upgrade to DALI 0.12.0 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 a new optimized implementation for CropMirrorNormalize for CPU and GPU.
  • Added the ability to read video files together with labels.

Fixed Issues

This DALI release includes the following fixes.

  • Resumed shipping the prebuilt version of the TensorFlow plugin as a part of nvidia-dali-tf-plugin (as a result of fixing a compatibility issue). With this fix, the user must first install nvidia-dali-tf-plugin package. When this package is installed, one of the prebuilt binaries of the TensorFlow plugin is selected. See Binary Installation.

Breaking Changes

  • Source code build now requires the compiler be at least C++14 capable.

Deprecated Features

  • Deprecated the following as noted below:
    • _run: use schedule_run instead
    • _share_outputs: use share_outputs instead, and
    • _release_outputs: use release_outputs instead.
  • Replaced HostDecoder and nvJPEGDecoder with generic ImageDecoder. ImageDecoder is the recommended function for the image decoding, and the replaced APIs HostDecoder and nvJPEGDecoder will be removed in the future.
  • NormalizePermute is replaced by the new CropMirrorNormalize.

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.