ReszkaMixes / README.md
reszka's picture
Update README.md
6a90319
|
raw
history blame
1.06 kB
---
license: creativeml-openrail-m
language:
- en
library_name: diffusers
pipeline_tag: text-to-image
tags:
- waifu-diffusion
---
# Model Card for Model ID
Anime models mix for stable diffusion
## Model Description
NEW GREAT MODELS MIX
Call it SkyrMix
This model is 50% weighted mix of:
https://huggingface.co/andite/anything-v4.0/blob/main/anything-v4.0-pruned.safetensors
and
https://huggingface.co/WarriorMama777/OrangeMixs/blob/main/AbyssOrangeMix2/AbyssOrangeMix2_nsfw.safetensors
use same settings
- **Developed by:** Reszka
- **Model type:** Anime/NSFW
- **License:** CreativeML OpenRAIL-M license
# Uses
Sampler: “DPM++ SDE Karras” is good
Steps: forTest: 12~ ,illustration: 20~
Clipskip: 1 or 2
Upscaler : Latenet (nearest-exact)
CFG Scale : 5 or 6 (4~8)
Denoise strength: 0.5 (0.45~0.6)
If you use 0.7~, the picture will change too much.
If below 0.45, Block noise occurs.
## Recommendations
Use bad_prompt_version2, bad-hands5 and bad-artist for better results
<details>
<summary> Click to expand </summary>
</details>