Metadata-Version: 2.4
Name: jax_loop_utils
Version: 0.0.14
Summary: Libraries for ML training loops in JAX.
Project-URL: Homepage, http://github.com/garymm/jax_loop_utils
Author-email: Gary Miguel <garymm@garymm.org>
License: Apache 2.0
License-File: LICENSE
Keywords: JAX,machine learning
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: License :: OSI Approved :: Apache Software License
Classifier: Programming Language :: Python
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.11
Requires-Dist: absl-py
Requires-Dist: etils[epath,epy]
Requires-Dist: jax>=0.4.36
Requires-Dist: numpy
Requires-Dist: wrapt
Provides-Extra: audio-video
Requires-Dist: av>=14.0; extra == 'audio-video'
Provides-Extra: mlflow
Requires-Dist: mlflow-skinny>=2.0; extra == 'mlflow'
Requires-Dist: pillow; extra == 'mlflow'
Provides-Extra: synopsis
Requires-Dist: chex; extra == 'synopsis'
Requires-Dist: flax; extra == 'synopsis'
Requires-Dist: ipykernel; extra == 'synopsis'
Requires-Dist: matplotlib; extra == 'synopsis'
Provides-Extra: tensorflow
Requires-Dist: tensorflow>=2.12; extra == 'tensorflow'
Provides-Extra: torch
Requires-Dist: torch>=2.0; extra == 'torch'
Description-Content-Type: text/markdown

# JAX Loop Utils

This repository contains common functionality for writing ML training loops in JAX.
The goal is to make trainings loops short and readable (but moving common tasks to
small libraries) without removing the flexibility required for research.

To get started, check out [this Notebook](./synopsis.ipynb), or just:

```sh
uv add jax-loop-utils
# or if you're not using UV
pip install jax-loop-utils
```

See [pyproject.toml](pyproject.toml) for the optional dependencies, which are
needed for specific metrid writers.

This started as a fork of [CLU](https://github.com/google/CommonLoopUtils).
See [CHANGELOG.md](./CHANGELOG.md) for more details on changes since the fork.
