LoRA FOR WEAK GRAPHICS CARDS.
🔺 I highly recommend using this model with the indigo furry Mix family of models — v120 hybrid, v110 realistic, Hardfurry
🔻Not recommended set above 0,1-0,5 strength for this LoRA,
At the moment, I do not recommend using this LoRA, as it is in the testing phase.
R0.5 Beta:
!!! LoRA is very good at the checkpoint Indigo furry mix v120 hybrid.
Recommendations:
• Positive prompt:
(LoRA), Raccoon21style:1, pastel tones, 8K, perfect face, good anatomy, best quality, realistic, cinematic, [write here whatever you want]
• Negative:
Bad anatomy, lowres, bad hands, boring e621, poor quality, multiple tails, dark, dark image, black shadows
Sampler: DPM++ 2M SDE Heun Karras (very recommend)
Clip skip: 1
Steps: 30 very good, >31 - 50 perfect.
Strength LoRA: 1, Never set more than 1.2, as this will make you porridge.
Model: Indigo furry mix v120 hybrid (Very recommend)
Resolution:
For portrait: 512×.... (not recommend high 892 px)
For square: 512×512 or 640×640
For landscape: .... × 512 (not recommended high 892 px)
FOR R1 BETA:
!!! VERY RECOMMEND FOR INDIGO FURRY MIX V110 REALISTIC
Also, I recommend setting the strength LoRA in the range 0.5-0.7. | 1.0 also good.
Good generation! I'm specifically making it based on SD1.5 because I have a GTX 1650. :))
Description
• a larger dataset
• More quality
