Metadata-Version: 2.1
Name: jaxfm
Version: 0.1.2a0
Summary: JAX-based implementation of Flow Matching
Author-email: Pierre-Louis Ruhlmann <pierre-louis.ruhlmann@inria.fr>
Project-URL: repository, https://github.com/pierrelouis1998/JaxFM/tree/main
Classifier: Intended Audience :: Science/Research
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: jax
Requires-Dist: numpy
Requires-Dist: tqdm
Requires-Dist: equinox
Requires-Dist: jaxtyping
Requires-Dist: matplotlib
Requires-Dist: optax

# JAXFM : A Python package for Flow Matching using JAX
<p align="center">
  <img src="logo.png" />
  <img src="logo2.png" />
</p>


## Description
A Python package implementing flow matching methods using jax. The code includes 
- `reflows.py` : Implementation of [1] 
- `conditional.py`: Implementation of [2]
Various test datasets are also available to test the above methods.

## Refrences

[1]: [(Liu et al.)](http://arxiv.org/abs/2209.03003) **Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow** 

[2]: [(Lipman et al.)](http://arxiv.org/abs/2210.02747) **Flow Matching for Generative Modeling**
