[Scene_Cast2]

Hasn't this been the case for a while? I vaguely remember using the optimal e8 packing to try to get denser product quantization for vector embeddings.

[corndoge]

She proved it

[wyndyl]

How did you apply e8 to vector embeddings?

[Scene_Cast2]

Chunk up your 128D embedding into 16 8D pieces, then quantize using the optimal packing points. Performs worse than k-means in my experience.

[bananaflag]

Why is this downvoted? I think it is sincere.