Database / Weaviate Vector database Interview questions
What is Rotational Quantization (RQ) in Weaviate?
RQ is Weaviate's newest and currently recommended compression technique, applying a fast pseudorandom rotation to each vector before quantizing it, which spreads information evenly across dimensions and allows high recall retention without any training phase or manual tuning.
vector_index_config=Configure.VectorIndex.hnsw( quantizer=Configure.VectorIndex.Quantizer.rq(bits=8) )
RQ is available in 8-bit and 1-bit variants: 8-bit RQ offers a moderate, roughly 4x compression ratio with very well-retained recall, while 1-bit RQ pushes toward roughly 32x compression as vector dimensionality increases, at some further recall cost, though still generally regarded as more robust and accurate than the older Binary Quantization (BQ) approach it's positioned to improve on. Because RQ needs no training set or tuning, Weaviate's own guidance suggests it as the sensible starting point for most HNSW (and now flat-index) compression needs, rather than defaulting to PQ or SQ, which both require a training phase to configure well.
More Related questions...
