Y
Hacker News
new
|
ask
|
show
|
jobs
by
diggan
606 days ago
> that support fast and lossless inference of 1.58-bit models on CPU (with NPU and GPU support coming next).