AnswerDotAI / AnswerDotAI/RAGatouille
Decouple Indexing and Encoding to support other late-interaction models
Open
enhancement
help wanted
Indexing expansion
- Dominant language
- Python
- Stars
- 4k
- Forks
- 276
- PR merge metrics
- No merged PRs in 30d
Description
Currently, Indexing is handled by upstream ColBERT completely, and only accepts ColBERT outputs.
We want to decouple the indexing stage and the document encoding stage, so that we can take in any ColBERT-like multi vector model as input.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.