🐻Moon Toon Mix Illustrious is an merge checkpoint that blends anime and cartoon elements with a clean, illustrative touch. It focuses on expressive, well-defined characters with soft color gradients and a high level of detail. This model works especially well for anime girls & boys, furry characters and digital-art styled scenes with a light toon aesthetic.
If you want to support me, post your first picture directly on this model page by clicking on "Add Post" below. And if you like this model, don't forget to 👍
📢Even a small⚡bid amount will help keep the model available for OnSite-Generation. Thank You!❤️
Recommended Usage:
Positive prompts: masterpiece, best quality, very aesthetic,
Negative prompts: worst quality, bad anatomy, lowres
Good Resolutions for SDXL:
832 × 1216 (recommended)
1024 × 1024
896 × 1152
1152 × 896
Sampler Suggestions:
Euler a, DPM++ 2M Karras
CFG 4–7, 20–40 steps
Black Images / VAE Precision
If your generations or inpainting results turn completely black, check your VAE precision settings.
Some setups may have issues when the VAE is running in FP16. Switching the VAE precision to FP32 can fix black or broken outputs.
If you encounter black images:
Check your current VAE precision.
Try switching VAE Precision from FP16 to FP32.
Reload the model/VAE after changing the setting if needed.
This fix was reported by @annoy1942, who encountered the issue while using InvokeAI.
Description
FAQ
Comments (27)
Well done! First tests feel really good. Gives a lot of versatility and remix potential. Guess I'm going to add it for my Toon base <3
Thank you! I'm so happy that it's working so well. You get tunnel vision after a while. But then to hear that it's working well for others and that they like it makes me very happy.🤘🐻🤘
@Moonbear_AIArt Absoluetly. Keep up your Great work 😍
This is one of the few models that works really well with VDM Solver. I don't know what you did, but it lets that highly creative sampler really flourish and show its strength.
Getting consistent results in sdnext, 1024x1024 with 40 steps using VDM Solver, karras, sigma .85. Refine to 1152x1600 with 20 steps using Heun at strength .5. I will post some samples later. This is easily my favorite model now.
Thank you so much! I'm really happy that it works so well and I'm looking forward to your samples. 🤘🐻🤘
@Moonbear_AIArt Done. Meanwhile, I tried Euler A, and it's awesome too. I just like the freaky in VDM Solver...
@Vera100 Oh, unfortunately, I don't see any images yet. Are they stuck somewhere in the rating analysis?
@Moonbear_AIArt Not sure... could be because it is my first post? Here is the link to the post, looks like it was approved, and it mentions your model. https://civitai.com/posts/19111848
I am also creating some images here on civitai, it's pretty good with decent samplers and a lot simpler for sure. I want this model to succeed and receive additional updates.
Works well with all kind of samplers really and all kind of resolutions... I like a solid model, versatile.
Thank you very much! 🤘🐻🐾
Really Nice Model. It's my new Favorite <3
Thank you ☺️
Amazing results, was worthy the bid
Thank you very much!⚡🐻🤘
Damn, the quality is impressive. Nice model and thanks for sharing it.
Endless thanks! 🐻🚀
👍
The checkpoint has returned to bid...
I hate bids bro this gotta be a joke
jeezosbeff360 Unfortunately not. That's the current ecosystem on CivitAI. I hate it just as much, but without your help, I can't cover the weekly costs on my own (the cut is around 8k-9k per week). I've already suggested a change and pointed out "promises" from four months ago in this article to improve this unfair system (https://civitai.com/articles/12904?highlight=1283492). We won't be able to get rid of it completely. Unfortunately, there are too many people who shamelessly exploit it.
Moonbear_AIArt I hate it when my computer tries to freeze every time there is a large number of comments on a civitai article.
DoubleSuger yes, caching and chunk loading are not good in Civitai -.-
Been away for a week, such a shame to find it in bid again
@ragnaroook Endless thanks! 🐻⚡🚀
I've been trying to use this model on InvokeAI with recommended settings but I can't make it work. It only generates pitch black things during inpainting. Any help?
I had never used InvokeAI before, but because of your comment I did a completely fresh install, imported the checkpoint and generated a test image using the default settings. Regular generation works fine for me, as shown in the screenshot. https://ibb.co/kg1d0dx9
I haven't tested inpainting specifically though, so I can't reproduce or rule out the issue you're seeing there. It may be related to some InvokeAI/inpainting-specific setting rather than the checkpoint itself.
Unfortunately I can't offer much help with InvokeAI-specific troubleshooting, since this was literally my first time using it.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.