ragit
| Crates.io | ragit |
| lib.rs | ragit |
| version | 0.1.0 |
| source | src |
| created_at | 2024-10-20 14:35:14.546414 |
| updated_at | 2024-10-20 14:35:14.546414 |
| description | git-like rag pipeline |
| homepage | |
| repository | |
| max_upload_size | |
| id | 1416249 |
| size | 286,033 |
(baehyunsol)
documentation
README
RAGIT
RAGIT (rag-it) is a git-like software that turns your local files into a knowledge-base. The main goal of this project is to make knowledge-bases easy to share.
It allows you to
- create and share knowledge-bases easily
- ask query on knowledge-bases
Why another RAG framework?
RAGIT is very different from the other RAG frameworks. The differences make RAGIT suitable for mid-size data files (a few 100MBs I guess), but not for very big source.
- It adds title and summary to every chunks. It makes AIs very easy to rerank chunks.
- It DOES NOT use vector DB. Not using vector DB makes it difficult to scale to million files, but instead, it's VERY easy to share your knowledge-bases with others.
- It calculates tf-idf score on every chunks. It must be fast enough for hundreds of thousands of chunks.