Skip to content

Latest commit

 

History

History
60 lines (43 loc) · 2.91 KB

README.md

File metadata and controls

60 lines (43 loc) · 2.91 KB

InstructLab Community Learning Guide

InstructLab crowd sources the process of tuning and improving models by collecting two types of data, knowledge and skills. These submissions are collected in a taxonomy of YAML files to be used in the synthetic data generation process.

Overview of the LAB alignment method. From Sudalairaj et al., 2 Mar 2024.

We accept contributions of both Skills and Knowledge to InstructLab.

Learning Topics

Skills

Knowledge

License Limitations

If you would like to contribute any third-party data to either the Skills or Knowledge taxonomies, you must ensure the license on the data is unrestricted for commercial use.

This applies to:

  • Data embedded in .md files as knowledge
  • Data offered as context in qna.yaml files for skills
  • Citing your sources in your attribution.txt file
  • Questions and answers sourced from elsewhere and used as qna.yaml submissions

For this project, unless the file says otherwise, or unless the attributed source provided in the file says otherwise, the relevant open source license is the Apache License, Version 2.0. All contributions that leverage third party content should either come from the public domain (e.g. out of copyright, or .gov sites) or be licensed with an open data license that does not restrict commercial use or the creation of derivative works, including the following license types:

  • CC0
  • CDLA-Permissive-2.0
  • CC-BY-4.0
  • Apache 2.0
  • MIT

Any third party content contributed to this project undergoes modifications in order to formulate it in the templated format required for submission to this project.

Works Cited on this Page