tveasey commented on PR #15903: URL: https://github.com/apache/lucene/pull/15903#issuecomment-4200729702
> rotation is the real secret sauce here One thing to note on rotations is that [block diagonal](https://www.elastic.co/search-labs/blog/robust-optimized-scalar-quantization) with random permutation performs basically as well as dense with block sizes of 64 x 64. This might be competitive with Hadamard given we can perform the 64d matmul extremely fast with SIMD. Regarding [this](https://github.com/apache/lucene/pull/15903#issuecomment-4195152417) comment I'm not sure we should mix up the choice of reranking representation and retrieval representation. There is also something a bit odd about these results: TQ-1 bit shows fairly consistent worse recall than BBQ (even without reranking). This makes me wonder if the accelerated distance calculation is just different. If we were to make the argument to use quantised representations for reranking on accuracy grounds (care is needed here) this suggests we should just use higher bit OSQ. -- 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]
