Scores with scalar quantization
What changes in Scores with scalar quantization compared with MongoDB and what your application has to do.
This difference is of the Different kind.
What MongoDB does
With quantization: "scalar", the score uses learned quantiles and changes from one build of the index to another.
What TeselaDB does
It scores with the full vectors; the order of the documents matches in what was measured.
What your application has to do
Nothing. The score does not change when the index is rebuilt.