dungba88 commented on issue #14007:
URL: https://github.com/apache/lucene/issues/14007#issuecomment-2505112049

   > Though, conceivably, the codec could quantize the vectors twice (int7, 
then to bit) and allows you to rerank bit quantized values with int7 quantized 
values...quantization all the way down
   
   I like this idea, it allows more granular trade-off between memory usage and 
recall. Although if it means we need to store all 3 levels (float32/int7/bit) 
in disk, or we need to sacrifice accuracy during merging by just relying on the 
intermediate int7.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to