rocAL documentation#
Deep learning applications require loading and pre-processing data efficiently to achieve high processing throughput. This requires creating efficient processing pipelines fully utilizing the underlying hardware capabilities. Some examples are load and decode data, and perform a variety of augmentations such as color-format conversions. Deep learning frameworks require supporting multiple data formats and augmentations to adapt to a variety of data-sets and models.
The ROCm Augmentation Library (rocAL) is designed to efficiently decode and process image and video pipelines from a variety of storage formats. These pipelines are programmable by the user using both C++ and Python APIs. rocAL is implemented in the HIP programming language and optimized for AMD’s latest discrete GPUs.
The code is open and hosted at: ROCm/rocAL
The rocAL documentation is structured as follows:
To contribute to the documentation refer to Contributing to ROCm Docs.
You can find licensing information on the Licensing page.