Open, Accelerated Machine Learning

MLCL, is an open-source replacement for your every-day machine learning tasks, with the key function of being crossplatform, efficient, and supports GPU acceleration with anything that supports OpenCL.

Purpose and Motivation

Machine learning can be challenging, particularly when running models on older devices or less common hardware, such as the RX580 graphics card or CPUs released in the past decade. MLCL addresses this gap by providing a user-friendly framework that facilitates machine learning tasks on a wide range of hardware configurations.

Features

• HDL Conversion
• Almost a 1-1 replacement for PyTorch (CNNs, RNNs, etc.)
• Automatic Accelerator detection