Verified: a year ago
SafeTensor
Type | |
Stats | 614 |
Reviews | (121) |
Published | Jul 22, 2023 |
Base Model | |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 6E6E9EA462 |
Mix of Yuzu, Hassaku and Sudachi 3 of my favorite models.
Model quite compatible with LoRas, the sample images are all made with loras, no Adetailer (after detailer) and no face restoration, 100% prompts.
used LoRa from More details and lycoris from Envybetterhands, to generate the sample images.
excellent for nsfw
Recommended settings:
Sampler: DPM++ 2M karras or DPM++ 2s Karras (they generate the faces very well and give a lot of detail).
clip skip: 2
steps: 25 to 30 (or more if possible)
CFG scale: 7 (this is what I used to generate the sample images but you can change it according to your needs)
do not use restore face.
share your generations in the comments
That's all. Enjoy it!