Comment by vectorrain

What are the advantages and potential challenges of combining dense vectors, sparse vectors, and full-text search in a hybrid retrieval method, as implemented in Infinity v0.2, and how does this approach compare to traditional vector search or other retrieval methods?