NVIDIA/DALI v1.39.0 on GitHub

Key Features and Enhancements

This DALI release includes the following key features and enhancements:

Added support for CUDA 12.5 (#5478).
Migrated fn.decoders.image* operators to use nvImageCodec as a decoding backend (#5470).
Improved error handling (#5466, #5494, #5486, #5491).

Fixed Issues

Fixed DALI TF plugin compatibility with TensorFlow 2.9 (#5499).
Fixed S3 fn.readers.file support for pad_last_batch=True (#5493).
Fixed a bug that resulted in long build times for some pipelines with enabled conditional execution (#5475).

Improvements

Add a mention of blogpost in Automatic Augmentation docs (#5508)
Removal of Python 3.8 notes from documentation (#5502)
Add default schema and use it in OpSpec argument queries. (#5500)
Add missing blocking argument documentation to the external source operator (#5501)
Trim line length in the documentation/examples for the new theme (#5479)
Refactoring in Pipeline, OpGraph and old Executor + name lookup improvement in old OpGraph and Pipeline. (#5495)
Improve error message about FFmpeg not being available (#5494)
Extend docs by adding info about @do_not_convert for NUMBA and Python ops (#5488)
New OpGraph (#5485)
Fix tests for sanitizer build (#5492)
Github comment acceptance formating table fix (#5490)
Remove image decoder memory padding from examples (#5484)
Adding git lfs as a compilation prerequisite (#5483)
Remove unused JIT workspace policy. (#5487)
Add a warning about pipeline definition being executed only once. (#5486)
Move to CUDA 12.5 (#5478)
Pin NPP version for CUDA 12 (#5480)
Reintroduce "Move old ImageDecoder to legacy module and make the nvImageCodec based ImageDecoder the default" (#5470)
Move to new, unified, NVIDIA sphinx theme (#5471)
Add DALI video plugin skeleton (#5328)
Don't initialize NVML when not setting affinity. (#5472)
Add MXNet deprecation message to the docs and plugin (#5465)
Add first-class check for nested datanodes in math/arithmetic ops. (#5466)

Bug Fixes

Fix DALI TF plugin incompatibility with TF 2.9 (#5499)
Coverity May 2024 (#5497)
Fix S3 FileReader when using repeated samples (pad_last_batch=True) (#5493)
Improve the video decoder errors (#5491)
Add extra rpath for prebuilt ffmpeg dependencies for video plugin (#5481)
Use dynamic programming in OpGraph::HasConsumersInOtherStage (#5475)

Breaking API changes

There are no breaking changes in this DALI release.

Deprecated features

DALI 1.39 is the final release that will support the MXNet integration.

Known issues:

The following operators: experimental.readers.fits, experimental.decoders.video, and experimental.inputs.video do not currently support checkpointing.
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.
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 plugin might not be compatible with TensorFlow versions 1.15.0 and later.
To use DALI with the TensorFlow version that does not have a prebuilt plugin binary shipped with DALI, make sure that the compiler that is used to build TensorFlow exists on the system during the plugin 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 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.

Binary builds

NOTE: DALI builds for CUDA 12 dynamically link the CUDA toolkit. To use DALI, install the latest CUDA toolkit.

CUDA 11.0 and CUDA 12.0 builds use CUDA toolkit enhanced compatibility. 
They are built with the latest CUDA 11.x/12.x toolkit respectively but they can run on the latest, 
stable CUDA 11.0/CUDA 12.0 capable drivers (450.80 or later and 525.60 or later respectively).
However, using the most recent driver may enable additional functionality. 
More details can be found in enhanced CUDA compatibility guide.

Install via pip for CUDA 12.0:
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/ nvidia-dali-cuda120==1.39.0
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/ nvidia-dali-tf-plugin-cuda120==1.39.0

or just:

pip install nvidia-dali-cuda120==1.39.0
pip install nvidia-dali-tf-plugin-cuda120==1.39.0

For CUDA 11:
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/ nvidia-dali-cuda110==1.39.0
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist/ nvidia-dali-tf-plugin-cuda110==1.39.0

or just:

pip install nvidia-dali-cuda110==1.39.0
pip install nvidia-dali-tf-plugin-cuda110==1.39.0

Or use direct download links (CUDA 12.0):

Or use direct download links (CUDA 11.0):

FFmpeg source code:

This software uses code of FFmpeg licensed under the LGPLv2.1 and its source can be downloaded here

Libsndfile source code:

https://developer.download.nvidia.com/compute/redist/nvidia-dali/libsndfile-1.2.2.tar.gz

NVIDIA/DALI v1.39.0 DALI v1.39.0 on GitHub