HackerTrans
TopNewTrendsCommentsPastAskShowJobs

demilich

no profile record

Submissions

Nemotron-4 340B

research.nvidia.com
3 points·by demilich·2 lata temu·0 comments

comments

demilich
·2 lata temu·discuss
https://github.com/infiniflow/infinity, dense vector + sparse vector + fulltext search(BM25) + late interact reranker(Colbert)
demilich
·2 lata temu·discuss
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.
demilich
·2 lata temu·discuss
Current RAG is more and more complex, will future LLM or New AI model overthrow the RAG?
demilich
·2 lata temu·discuss
Infinity supports HNSW vector index.
demilich
·2 lata temu·discuss
Check https://github.com/infiniflow/infinity which combines vector search and full-text search providing extremely fast search performance.
demilich
·2 lata temu·discuss
Try RAPTOR: https://arxiv.org/html/2401.18059v1

An implementation: github.com/infiniflow/ragflow
demilich
·2 lata temu·discuss
Agreed
demilich
·2 lata temu·discuss
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?
demilich
·2 lata temu·discuss
Use C++20 modules, take a look at this project: https://github.com/infiniflow/infinity
demilich
·2 lata temu·discuss
Cool! This is really helpful.
demilich
·2 lata temu·discuss
Each project has its own detailed requirements and scenarios, and we cannot demand that each project use same library to implement similar functions
demilich
·2 lata temu·discuss
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.