I'd like ccc to be able to run ColBert-style embedding models such as lightonai/LateOn-Code using the PyLate framework locally.
Multi-Vector Small Embedding models are lightweight to run locally, yet powerful for retrieval accuracy when used with the Late Chunking technique.
Example
https://github.com/lightonai/next-plaid
I'd like
cccto be able to run ColBert-style embedding models such aslightonai/LateOn-Codeusing the PyLate framework locally.Multi-Vector Small Embedding models are lightweight to run locally, yet powerful for retrieval accuracy when used with the Late Chunking technique.
Example
https://github.com/lightonai/next-plaid