You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
All our vector benchmarks use float. It would be good to benchmark byte linearly quantized along side of float.
We need to measure indexing & search speed for byte vs float.
I think the thing to do here is take the so_vectors or dense_vectors data set and linearly quantize all of it into byte values and upload a new copy that rally can use.
The text was updated successfully, but these errors were encountered:
All our vector benchmarks use
float
. It would be good to benchmarkbyte
linearly quantized along side offloat
.We need to measure indexing & search speed for byte vs float.
I think the thing to do here is take the
so_vectors
ordense_vectors
data set and linearly quantize all of it intobyte
values and upload a new copy that rally can use.The text was updated successfully, but these errors were encountered: