Skip to content

A high-performance RAG system for PDFs using multi-vector embeddings (ColPali / ColQwen / ColSmol) with vector search in Qdrant, prefetch optimization, and reranking for improved relevance. Designed for speed, accuracy, and scalability, this system is ideal for building intelligent search, document understanding, and QA applications.

Notifications You must be signed in to change notification settings

logxdx/contextualized-late-interation-with-pdfs

About

A high-performance RAG system for PDFs using multi-vector embeddings (ColPali / ColQwen / ColSmol) with vector search in Qdrant, prefetch optimization, and reranking for improved relevance. Designed for speed, accuracy, and scalability, this system is ideal for building intelligent search, document understanding, and QA applications.

Topics

Resources

Stars

Watchers

Forks

Languages