South Park Style (LORA)版本v2.0 (ID: 94964)

(v2.0 update): Fine-tuned the model some more. I also added side views. Insert side view in your prompt, and just in case if it doesn't work the first time, be sure to use ((side view)) to get more weight to your image.

I went down to South Park and created a LoRA out of the art style!

Recommended settings:

If you want an adult version of the South Park characters, type in the words "tall" and/or "adult", and the model should be giving you the adult model more accurately.

Sampler: DPM++ 2M Karras, or Euler a

LoRA weight: ~0.5-0.6

CFG scale: ~6-7

If you are doing hires. fix, I recommend the denoising weight of ~0.5-0.7. (And I also recommend you increase the steps for more accuracy; up to at least 40)

It's optional, but just in case, if your images result in a realistic photo, just add the words:

cartoon, south park style

Checkpoints: I recommend Deliberate, since I mostly tested the images on it.

Negative prompt that works well:

deformed, bad anatomy, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, disgusting, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands, blurry, ((mutated hands and fingers)), duplicate images, duplicate, (deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, EasyNegative, ng_deepnegative_v1_75t

DeepNegative and EasyNegative are very recommended.

This is only the first version, and I will iron out any problems in the next ones.

Suggestions and feedback are welcome! I hope you enjoy

描述:

(v2.0 update): Fine-tuned the model some more. I also added side views. Insert side view" in your prompt, and just in case if it doesn't work the first time, be sure to use ((side view)) to get more weight to your image.

训练词语: south park style