DALI Release 1.16.0

Using DALI 1.16.0

To upgrade to DALI 1.16.0 from an older version of DALI, follow the installation and usage information in the DALI User Guide.

Note: The internal DALI C++ API used for operator’s implementation, and the C++ API that enables using DALI as a library from native code, is 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 GPU non-silent region detection operator (#3944, #4001).
  • Added experimental support for the eager execution of stateful and arithmetic operators (#4016, #3952, #3969, #3990).
  • Added an antialias flag to the resize operator for improved control over the resampling mode that is used (#4032).
  • Added experimental support for custom GPU Numba operators (#3891, #3998, #4006, #4013).
  • Added support for the processing video and the handling of temporal arguments to color-manipulation operators and affine transform operators (#3937, #3946, #3917).

Fixed Issues

Here are the fixed issues in this release:
  • Fixed the DALI + PyTorch Lightning iterator issue where subsequent epochs were terminating too early (#3923, #4048).
  • Fixed scalars handling by the readers.tfrecord operator (#4024).
  • Fixed the variable batch size handling by the crop and coord_transform operators (#4045, #3958).

Breaking Changes

This DALI release includes the following breaking changes:
  • The shape of scalars that are read by the readers.tfrecord operator is now () instead of (1,).
  • For the cubic and linear interpolation modes, by default, the resize operator applies the antialiasing filter.

    The antialiasing can be disabled by using the antialias flag.

Deprecated Features

The following feature has been deprecated in this release:
  • The triangular interpolation for the resize operator is identical to the linear interpolation with antialiasing enabled, so it has been deprecated.

Known Issues

This DALI release includes the following known issues:

  • 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 frequency that is less than 10-15 frames, the returned frames might be out of sync.

  • The experimental VideoReaderDecoder does not support open GOP.

    It will not report an error and might produce invalid frames. VideoReader uses a heuristic approach to detect open GOP and should work in most common cases.

  • The DALI TensorFlow plug-in might not be compatible with TensorFlow versions 1.15.0 and later.

    To use DALI with the TensorFlow version that does not have the prebuilt plug-in binary that is shipped with DALI, ensure that the compiler that is used to build TensorFlow exists on the system during the plug-in installation. (Depending on the particular version, you can use GCC 4.8.4, GCC 4.8.5, or GCC 5.4.)

  • In experimental debug and eager modes, the GPU external source is not properly synchronized with DALI internal streams.

    As a workaround, you can manually synchronize the device before returning the data from the callback.

  • Due to some known issues with meltdown/spectra mitigations and DALI, DALI shows the best performance when running in Docker with escalated privileges, for example:
    • privileged=yes in Extra Settings for AWS data points
    • --privileged or --security-opt seccomp=unconfined for bare Docker