fastax

Overview

fastax is a Deep Learning Library built on top of Google's JAX that tries to facilitate doing research by providing a combination of JAX's versatility with implementations of the most used neural networks layers and utilities.

More concrete information and naming

fastax is both a combination of fastai and JAX and fastai and stax:

• It is built on top of JAX and most of its original code comes from its Neural Network mini-library stax
• One of the long term goals of this library is implementing Jeremy Howard's fastai course "Deep Learning From the Foundations"

For more information visit https://github.com/joaogui1/fastax