• Ani-Toon •
This Lora was specially trained to create a mix of different cartoon styles but with the quality and detail of a common anime style.
• Suggestions for a better use •
It is recommended to use this LORA with a high weight between 0.6 and 0.9 approximately. If you want to use it together with other quality indicator models, it is advisable to leave the latter with very low weights around 0.3 and 0.6 so as not to distort the cartoon aspect that characterizes this project.
Using ControlNet in its "Lineart Anime" version also significantly improves image generation.
• Generation Specifications •
Sampling Method: DMP++ SDE Karras
Sampling Steps: 25 - 30
Size: 384 x 512
Hires. Fix: Enable
Resize: 768 x 1024
Upscaler: R-ESRGAN 4x+ Anime6B
Hires Steps: 0
Denoising: 0.6
CFG Scale: 7 - 9