34 ms·
How do you weight results between vector search and bm25? Do you fall back to bm25 when vector similarity is below a threshold, or maybe you tweak the weights b
by drittich 2y ago
How do you weight results between vector search and bm25? Do you fall back to bm25 when vector similarity is below a threshold, or maybe you tweak the weights by hand for each data set?
- gillesjacobs 2y agoThe algorithm I use to get a final ranking from multiple rankings is called "reciprocal ranked fusion". I use the implementation described here: https://docs.llamaindex.ai/en/stable/examples/low_level/fusion_retriever/#step-3-perform-fusion https://docs.llamaindex.ai/en/stable/examples/low_level/fusi... Which is the implementation from the original paper.
- drittich 2y agoThanks, much appreciated!