Y
Hacker News
new
|
ask
|
show
|
jobs
by
rohansood15
92 days ago
The paper is about vector quantization, which affects KV cache not model weights/sizes.