Fantastic video! I loved how clearly you explained building the ultimate hybrid search with Qdrant. The step-by-step guide and practical examples made it so easy to follow. Can't wait to implement this in my projects! Any tips on optimizing performance for large datasets? Thanks for sharing!
can i use bm25 in combination with open ai embedding (tekt-embedding-3-smal)? I intend to put about 300 GB of pdf books in qdrant?is there a problem if i omit late interaction embedding?
I think the best 2 options were dense+sparse with RRF and dense+sparse with Late interaction reranking(As we're only reranking the final result, it is faster than searching the whole late interaction vectordb).
Great presentation! Nice to see the different search options in practice 🙏🏾
these new qdrant features look insane, great webinar, I have to test all this myself, really inspiring, thanks!
Fantastic video! I loved how clearly you explained building the ultimate hybrid search with Qdrant. The step-by-step guide and practical examples made it so easy to follow. Can't wait to implement this in my projects! Any tips on optimizing performance for large datasets? Thanks for sharing!
More of these!
can i use bm25 in combination with open ai embedding (tekt-embedding-3-smal)? I intend to put about 300 GB of pdf books in qdrant?is there a problem if i omit late interaction embedding?
I think the best 2 options were dense+sparse with RRF and dense+sparse with Late interaction reranking(As we're only reranking the final result, it is faster than searching the whole late interaction vectordb).
Kacper, is there a way to reach out to you?