jstmm
Joined 3 years ago
Karma
5
Submissions
5
Member since
Dec 8
Joined 3 years ago
On the contrary, in the BitNet b1.56 paper [0] the authors report their 7b model has 2.9x reduced inference latency.
It's not clear to me what's happening here. Can someone explain why the 1/2bit HQQ/HQQ+ models are so much slower than the BitNet b1.56 models?
https://www.coursera.org/specializations/modernrobotics#cour...