Re:Microsoft BitNet: 100B Param 1-Bit model for local CPUs
by naasking on Wednesday March 11, 2026 @11:07AM
(#47336625)
What's unusual about it? It seems pretty standard to train small models to validate an approach, and then show that training scales with model size to 8B to 14B parameter models, which is what they did.
[ Parent
| on HN ]
There are no comments on this comment yet.