Vae and Clip are all built in, so no need to Set Last Clip or a Vae. It is a V_Prediction model instead of the normal ESP models. Basic workflow and an advanced workflow included.
EclecticEuphoria_Noobai_vpred_C
Images Created with EclecticEuphoria_Noobai_vpred_C Civitai Lots and lots of images to see what the model can all do
It is a model merge and clip merge to introduce a bit more natural language. A nice balance if you only use Danbooru tags or if you only use Natural Language. This model likes to go NSFW easily workflow included
Works with various samplers but I liked unipc/beta - cfg 5 - 25-30 steps | works with and without skip clip, only time I ever used skip clip is if I wanted to alter the pic a bit more, otherwise I don't use skip clip
Initial Model Release
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support