COCStyle_XL

CoC-Style LoRA

Hello everyone,

I'm excited to share my very first trained LoRA model with you all!

This model is designed to generate images in a CoC style. It currently performs quite well for generating characters.

This is a work in progress, and I plan to continue iterating on it to improve its performance and versatility.

Hope you enjoy it! Please feel free to share your creations and any feedback you might have. I'd love to see what you make!

Usage Tips:

• Prompting:

  This LoRA supports both natural language and tag-style prompts. However, for the best and most consistent results, using individual tags and short phrases is highly recommended.

• Base Model:

  sd_xl_base_1.0

  URL：https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/tree/main

• LoRA Weight:

  0.7 ~ 1.0

• CFG Scale:

  7 ~ 10

• Resolution:

  1024x1024

• Samplers:

  Recommended: DPM2, DPM++ 2M,Euler

• Scheduler: Karras

• Steps: 30

• A Note on DDIM: You might notice some noise in the background when using the DDIM sampler. However, the quality of the main character is often still quite good, so feel free to give it a try if you don't mind the trade-off.

• Hires. Fix Settings:

  Upscaler: R-ESRGAN 4x+

  Denoising strength: 0.35 ~ 0.45