infini-gram 2.1.0

A Python package for infini-gram

infini-gram

To use this package, please check out its documentation here: https://infini-gram.io/pkg_doc

To learn more about infini-gram:

• Paper: https://arxiv.org/abs/2401.17377
• Project Home: https://infini-gram.io
• Web Interface: https://infini-gram.io/demo
• API Endpoint: https://infini-gram.io/api_doc

License

This package is licensed under the UW Academic Software License. Use by universities and non-profit institutions is allowed. Commercial use is not allowed. A copy of the license is enclosed with the package distribution.

The suffix array implementation is adapted from Lee et al. (2021), which is distributed under Apache-2.0.

Citation

If you find infini-gram useful, please kindly cite our paper:

@article{Liu2024InfiniGram,
  title={Infini-gram: Scaling Unbounded n-gram Language Models to a Trillion Tokens},
  author={Liu, Jiacheng and Min, Sewon and Zettlemoyer, Luke and Choi, Yejin and Hajishirzi, Hannaneh},
  journal={arXiv preprint arXiv:2401.17377},
  year={2024}
}