search

PrithivirajDamodaran / FlashRank

Lite & Super-fast re-ranking for your search & retrieval pipelines. Supports SoTA Listwise and Pairwise reranking based on LLMs and cross-encoders and more. Created by Prithivi Da, open for PRs & Collaborations.
Apache License 2.0
441 stars 37 forks source link

Update `README.md` to mention any models trained on copyrighted data #7

Closed chrispy-snps closed 4 months ago

chrispy-snps commented 4 months ago

This is an enhancement request.

Could you please update the README.md file to indicate whether any of the models were trained on copyrighted data, and if so, which ones? Thanks!

PrithivirajDamodaran commented 4 months ago

Out of scope: As readme states Flashrank brings existing SoTA ranking models Fx cross encoders for optimal reranking, it doesn't train or own models. Original Model owners of each model have written detailed model cards for all the models in huggingface. The models are Apache 2.0 licensed but you should Look up the card and / or contact the model owners as they are the right people to answer.