Skip to content

Latest commit

 

History

History
5 lines (3 loc) · 296 Bytes

README.md

File metadata and controls

5 lines (3 loc) · 296 Bytes

Script for using large language models to assign quality ratings to training examples in bulk.

Has two backends, vLLM and exllama. Outputs ratings to a shelve file.

Currently pretty specifically tailored to my workflow. Feel free to reach out to me with any questions, suggestions, or feedback.