Flax Models

A collection of pretrained models in Flax.

About

The goal of this project is to make current deep learning models more easily available for the awesome Jax/Flax ecosystem.

Models

• GPT2 [model]
• StyleGAN2 [model] [training]
• ResNet{18, 34, 50, 101, 152} [model] [training]
• VGG{16, 19} [model] [training]
• FewShotGanAdaption [model] [training]

Installation

You will need Python 3.7 or later.

1. For GPU usage, follow the Jax installation with CUDA.
2. Then install:

   > pip install --upgrade git+https://github.com/matthias-wright/flaxmodels.git

For CPU-only you can skip step 1.

Documentation

The documentation for the models can be found here.

Checkpoints

The checkpoints are taken from the repositories that are referenced on the model pages. The processing steps and the format of the checkpoints are documented here.

Testing

To run the tests, pytest needs to be installed.

> git clone https://github.com/matthias-wright/flaxmodels.git
> cd flaxmodels
> python -m pytest tests/

See here for an explanation of the testing strategy.

Acknowledgments

Thank you to the developers of Jax and Flax. The title image is a photograph of a flax flower, kindly made available by Marta Matyszczyk.

License

Each model has an individual license.