demilich

https://github.com/infiniflow/infinity, dense vector + sparse vector + fulltext search(BM25) + late interact reranker(Colbert)

GraphRAG is to parse data to create a KG and retrieval the information from KB.

RAGFlow is to create a Graph workflow to solve multi-hop question-answering issue.

Current RAG is more and more complex, will future LLM or New AI model overthrow the RAG?

Infinity supports HNSW vector index.

Check https://github.com/infiniflow/infinity which combines vector search and full-text search providing extremely fast search performance.

Try RAPTOR: https://arxiv.org/html/2401.18059v1

An implementation: github.com/infiniflow/ragflow

Agreed

RAGFlow (github.com/infiniflow/ragflow) use OCR/layout recognition/TSR(table structure recognition) to understand the document structure and context. Is there any difference between RAGFlow and ZenDB?

Use C++20 modules, take a look at this project: https://github.com/infiniflow/infinity

Cool! This is really helpful.

Each project has its own detailed requirements and scenarios, and we cannot demand that each project use same library to implement similar functions

The recognition of file layout to parse file content is indeed a very innovative idea. Compared to many open-source projects we have seen before, this provides us with new ideas for using RAG to solve problems in the future. I hope the author of this project will continue to update it, so that more people can benefit from it.

Well, I suppose its website have answered the similar question. From my point of view, FAISS and ANNOY are algorithm libraries which can be integrated into any application, but can't be used directly. Meanwhile, Milvus is a server can provide ANNS services independently, which already integrated many ANNS libraryies like FAISS and ANNOY. On the other hand, neighter FAISS nor ANNOY can solve all scenarios which have different vectors data size and distribution. As a server, Milvus can choose right algorithm to solve the problem.

Good job, excited!