This is a simple, but extremely functional, replacement for SDXL's cartoon keyword and is now V2. It has a different feel, look, and general style to it than V1.
There was a problem with the header that did not allow merging, and I have fixed that with the original file replaced with the corrected one.
Feeling generous? https://www.buymeacoffee.com/generalawareness
Description
Trained on a new Vintage cartoon style for this version two, and switch over to LoRA.
FAQ
Comments (6)
Just commenting to say this is very needed, there are a lot of overtrained / undertrained concepts in the base model. Thanks!
I did notice it made things so much better (take for instance the knives) than the base model that it surprised me (look at the blood, and while it is cartoonish it is so greatly superior to the base model, as someone told me).
I am glad you are enjoying it.
@general_awareness I've noticed there are two sort of over / under trained "sand traps" in the base SDXL model, one of them is very very noisy, like a fine noise over everything. The other is a particular line style and color scheme which is everywhere "illustration" is prompted. Seems wildly overtrained. I'm sure there are lots of different things like this people will uncover with the base model! Pretty sure it was left to the community to figure them out.
@revolved A lot of issues were introduced with 1.0 vs 0.9 as they did with 2.1 vs 2.0. Baking in the vae was a horrible idea.
This LoRA does a fantastic job with all kinds of comic book/graphic novel styles. Thanks for sharing it!
You are very welcome.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.