Sign In

Fluff'n'Pudge [mixed model]

25
397
4
Verified:
SafeTensor
Type
Checkpoint Merge
Stats
105
0
Reviews
Published
Jan 12, 2024
Base Model
SD 1.5
Usage Tips
Clip Skip: 1
Hash
AutoV2
955899F069

A merge model based on FluffyRock and the BiggerGirls-photorealism model.

UPDATE 1

  • Version v2: model for general use.
    Recipe: (((Deli-YMix30+FTas30+Zack20+BBro20-SD15*50+FluffyRock*50)*50+BiggerGirls_photorealism*50)*50+FluffyRock*50)
    Additional styles pack available here: https://drive.google.com/file/d/1PIU1Buaa63dQn3Z4nTQI0FQ2DFzZKnH2/view?usp=sharing
    Style examples added in gallery below.

  • Version v3: model for more "realistic" pictures, lower influence of specific artistic style.
    Recipe: (((BB95v11*50+BiggerGirls_photorealism*50)*50+(BB95v11*50+YiffyMix*50)*50)*80+FNPv2*20)

UPDATE 2

UPDATE 3

  • Version v.3 is online! New merge recipe: (((BB95v12*70+BiggerGirls_photorealism*50)*50+(YiffyMix*70+SyModel*30)*50)*80+FNPv3*20)

  • The model is more sensitive to both styles and authors, and allows to generate a wider arrays of pictures. It's also quite good at bigger body frames without using LORAs (just use tags like 'morbidly obese').

  • The CFG Rescale tool is not required but it's highly suggested, with values higher than 0.7. Keep CLIP Skip at 1 for best results.