← Back to context

Comment by liuliu

3 hours ago

What are you talking about? The model is native NVFP4, why you run it at any precision higher than that?

0 comments

liuliu

Reply

No comments yet

Contribute on Hacker News ↗