Toon-Tacular Qwen LoRA for Stable Diffusion
A LoRA (Low-Rank Adaptation) model trained on 70 curated images to give Stable Diffusion images a mid-to-late 90s/early 2000s cartoon-like style.
Why it matters
This LoRA provides a convenient way to add a unique cartoon-inspired style to Stable Diffusion-generated images, expanding the creative possibilities for artists and designers.
Key Points
- 1Trained on 70 images to create a consistent cartoon-inspired style
- 2Tested with Qwen Image 2512 model, has some weaknesses with busy backgrounds and smaller faces
- 3Trigger word is 't00n' but not necessary, 'animation' or 'cartoon' also work
- 4Recommended strength is 0.7-0.9, Euler/Beta sampler/scheduler
Details
The Toon-Tacular Qwen LoRA was trained using the ai-toolkit and tested in the Comfy UI. It was trained on a curated dataset of 70 images to create a consistent cartoon-inspired style reminiscent of mid-to-late 90s and early 2000s cartoons. The dataset was regularized using an edit model to upscale and unify the style. The LoRA was trained with the fp16 version of the Qwen Image 2512 model, and while it's not perfect, it generally maintains the cartoon style consistently. The LoRA has some weaknesses with overly busy backgrounds and smaller faces, so users are advised to be strategic in their prompting to get the best results.
No comments yet
Be the first to comment