ROCm Performance Primitives documentation

ROCm Performance Primitives (RPP) is a computer vision library for AMD CPUs and GPUs that have a HIP or CPU backend.

RPP implements image, voxel, and audio augmentations used in deep learning training pipelines. The same primitives run on the HOST CPU backend and the HIP GPU backend.

The RPP project is located in ROCm/rocm-libraries.

Install

• Install RPP

• Build from source

How to

• Run a tensor augmentation

Examples

• RPP examples

Reference

• RPP environment variables

• Supported RPP functionalities and variants

• RPP functionality and variant example outputs

• RPP API reference guide

To contribute to the documentation refer to Contributing to ROCm.

You can find licensing information on the Licensing page.