Re:Some notes on the Stable Diffusion safety filter
by xtagon on Friday November 18, 2022 @09:50PM
(#33666285)
This (training a model with no NSFW content) would be preferable to me. No false positives to worry about. People who do want to generate NSFW stuff can fine-tune or train their own model, nobody owes that functionality to them in freely available ones.
[ Parent
| on HN ]
There are no comments on this comment yet.