Mixed Precision Quantization on mlx comes with TurboQuant implementation

(twitter.com)

2 points | by jsilence 13 hours ago ago

1 comments