This is a simple utility LoRA for <a target="_blank" rel="ugc" href="https://civarchive.com/models/123841/comiccraft">ComicCraft</a>. Since ComicCraft focuses on comic and fantasy styles, they often have too many details to produce something that is closer to a cartoon or animation style and controlling it with just prompt can take some time.What this LoRA does is to simplify images, mainly by flattening colors (less gradients), reduce ink shadow effects and reduce the amount and strength of lines (both outlines and internal details), so it can be closer to that animation style.It works well for this purpose with weights between 0.75 and 1, but higher values will keep simplifying the images. Negative values will technically add more details, but will become noisy.It has been tested with other models and it works well too. For example used on realistic models, results become more semi-realistic illustrations and you can use higher weights to obtain results that are more like cartoons.The tags from the base ComicCraft should still be available, but will also be simplified.Regarding the showcase images, I didn't use Hires fix and tried to use examples from other similar models so it can be used for comparison.

CartoonCraft

CartoonCraftLora_V1.safetensors

v1.0

This is a simple utility LoRA for <a target="_blank" rel="ugc" href="https://civitai.com/models/123841/comiccraft">ComicCraft</a>. Since ComicCraft focuses on comic and fantasy styles, they often have too many details to produce something that is closer to a cartoon or animation style and controlling it with just prompt can take some time.What this LoRA does is to simplify images, mainly by flattening colors (less gradients), reduce ink shadow effects and reduce the amount and strength of lines (both outlines and internal details), so it can be closer to that animation style.It works well for this purpose with weights between 0.75 and 1, but higher values will keep simplifying the images. Negative values will technically add more details, but will become noisy.It has been tested with other models and it works well too. For example used on realistic models, results become more semi-realistic illustrations and you can use higher weights to obtain results that are more like cartoons.The tags from the base ComicCraft should still be available, but will also be simplified.Regarding the showcase images, I didn't use Hires fix and tried to use examples from other similar models so it can be used for comparison.

Files with SHA256: 0a28547e40954d04496886f9f4323ef0aac1c13e5e09c3b73ff31f2be1625f41