Package org.apache.lucene.util.quantization
Provides quantization methods for scaling vector values to smaller data types and possibly fewer
dimensions
-
Interface Summary Interface Description QuantizedVectorsReader Quantized vector readerRandomAccessQuantizedByteVectorValues Random access values forbyte[], but also includes accessing the score correction constant for the current vector in the buffer.ScalarQuantizedVectorSimilarity Calculates and adjust the scores correctly for quantized vectors given the scalar quantization parametersScalarQuantizedVectorSimilarity.ByteVectorComparator Compares two byte vectors -
Class Summary Class Description QuantizedByteVectorValues A version ofByteVectorValues, but additionally retrieving score correction offset for Scalar quantization scores.ScalarQuantizedVectorSimilarity.DotProduct Calculates dot product on quantized vectors, applying the appropriate correctionsScalarQuantizedVectorSimilarity.Euclidean Calculates euclidean distance on quantized vectors, applying the appropriate correctionsScalarQuantizedVectorSimilarity.MaximumInnerProduct Calculates max inner product on quantized vectors, applying the appropriate correctionsScalarQuantizer Will scalar quantize float vectors into `int8` byte values.