Metadata-Version: 2.4
Name: gfnx
Version: 0.0.1
Summary: Environments for Generative Flow Networks in JAX
Project-URL: Homepage, https://github.com/d-tiapkin/gfnx
Project-URL: Issues, https://github.com/d-tiapkin/gfnx/issues
Author: Daniil Tiapkin, Arten Agarkov, Nikita Morozov, Ian Maksimov, Timofei Gritsaev, Sergey Samsonov
License-Expression: MIT
License-File: LICENCE
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.10
Requires-Dist: chex
Requires-Dist: flashbax
Requires-Dist: jax-tqdm
Requires-Dist: jax>=0.4.27.0
Requires-Dist: jaxlib>=0.4.27.0
Requires-Dist: jaxtyping
Requires-Dist: matplotlib
Requires-Dist: networkx
Requires-Dist: numpy==1.26.4
Requires-Dist: omegaconf
Provides-Extra: baselines
Requires-Dist: equinox; extra == 'baselines'
Requires-Dist: flashbax; extra == 'baselines'
Requires-Dist: hydra-core; extra == 'baselines'
Requires-Dist: jax-tqdm; extra == 'baselines'
Requires-Dist: jraph; extra == 'baselines'
Requires-Dist: optax; extra == 'baselines'
Requires-Dist: orbax-checkpoint; extra == 'baselines'
Requires-Dist: pillow; extra == 'baselines'
Requires-Dist: wandb; extra == 'baselines'
Provides-Extra: tests
Requires-Dist: equinox; extra == 'tests'
Requires-Dist: pytest; extra == 'tests'
Description-Content-Type: text/markdown

# gfnx: Generative Flow Networks in jaX

`gfnx` is a JAX-native toolkit for building and studying Generative Flow Networks (GFlowNets). It brings together a collection of benchmark environments and reproducible baselines so you can iterate quickly on new ideas.

## Highlights

- End-to-end JAX implementations of GFlowNet building blocks (environments, reward modules, networks, and metrics).
- Ready-to-run baseline agents inspired by the [CleanRL](https://github.com/vwxyzjn/cleanrl) style of concise single-file experiments.
- Utilities for logging, checkpointing, and evaluation that make it easy to compare runs and extend the library with new research code.

## Installation

### Requirements

- Python 3.10 or newer.
- A working JAX installation. CPU works out of the box; for GPU/TPU accelerators follow the [official JAX installation guide](https://jax.readthedocs.io/en/latest/installation.html).

### Install the latest release

```
pip install git+https://github.com/d-tiapkin/gfnx.git
```

Verify the install with:

```
python -c "import gfnx; print('gfnx import OK')"
```

### Develop locally and run baselines

```
git clone https://github.com/d-tiapkin/gfnx.git
cd gfnx
pip install -e .[baselines]
```

The editable install keeps your local changes in sync with the Python package, while the optional `baselines` extra pulls in the dependencies required by the reference training scripts. As in with `CleanRL` ideology, the baselines are not supposed to be imported, they serve only as a reference implementation.

## Quickstart

Kick off a short training run of Detailed Balance in the Hypergrid environment:

```
python baselines/db_hypergrid.py num_train_steps=1_000 logging.tqdm_print_rate=100
```

The script is powered by Hydra, so you can override any configuration value on the command line (for example, picking another logging backend or playing with hyperparameters of the method). Baseline outputs, checkpoints, and Hydra logs default to `tmp/<date>/<time>/`; point the `logging.log_dir` or `logging.checkpoint_dir` fields to custom paths when running longer experiments.

## Support

- Open an issue on [GitHub](https://github.com/d-tiapkin/gfnx/issues) for bugs or feature requests.
- Start a discussion or reach out via pull requests if you would like to contribute improvements. Contributions with reproducible experiments and clear documentation get merged fastest.

## License

`gfnx` is released under the [MIT License](https://github.com/d-tiapkin/gfnx/blob/main/LICENCE). Feel free to use it in academic and commercial projects; please attribute the original authors when you publish results built on this codebase.

## Influences

`gfnx` stands on the shoulders of several excellent open-source projects:

- [torchgfn](https://torchgfn.readthedocs.io/en/latest/) – PyTorch-first GFlowNet library that shaped our environment design.
- [CleanRL](https://github.com/vwxyzjn/cleanrl) – taught us the value of single-file baselines and reproducible experiment configs.
- [purejaxrl](https://github.com/luchris429/purejaxrl/tree/main) and [JaxMARL](https://github.com/FLAIROx/JaxMARL/tree/main/jaxmarl) – reference points for idiomatic, accelerator-ready JAX reinforcement learning code.
