Cartoon 2D animation is a custom merge checkpoint designed for expressive, stylized 2D illustrations with flexible rendering modes. Whether you're aiming for soft painterly portraits, vibrant cartoon sketches, or detailed comic‑style compositions, this model adapts to your prompt structure and delivers consistent, emotionally rich results.
It excels at character‑driven storytelling, fantasy scenes, cozy interiors, sci‑fi drama, and visual novel aesthetics — all with clean color harmony and confident anatomy.
🤌🏻 Best for:
2D anime/cartoon portraits with expressive emotions and clean shading
Narrative scenes — from magical forests and moonlit beaches to futuristic corridors and quiet rooms
Consistent character design across multiple renders and angles
✨ Core Features:
Soft lighting and smooth shading without heavy depth or specular highlights
Illustration-style rendering with readable silhouettes and gentle gradients
Harmonious color palettes with controlled saturation
Confident anatomy and stylized proportions
Works well with both cozy and dramatic compositions
☝ Tips:
This model responds dynamically to your prompt's opening tags. Here's how to guide its style:
masterpiece, best quality
Soft 2D realism/Minimal cartoon stylization, smooth gradients, cinematic lightingtoon, flat colors, highres, amazing quality, 4k, high resolution, ultra-detailed, newest
Flat cartoon/Clean outlines, cel-style shading, vibrant and playfulmasterpiece, best quality, amazing quality, 4k, high resolution, ultra-detailed, comicbook style
Graphic stylization/Mid-level cartoon look with strong details and confident formsmasterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres, newest, cartoon, toon
Aesthetic hybrid/Stylized but detailed, with atmospheric depth and emotional tonemasterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres, newest, cartoon, toon, comicbook style
High-resolution clarity/Sharp rendering, rich textures, and modern polish
ADetailers: face_yolov8m, person_yolov8n-seg, Eyes, hand_yolov8n, face_yolov8n, full_eyes_detect_v1, Anzhc%20Face%20-seg, Anzhc%20Eyes%20-seg-hd, PitHandDetailer-v1b-seg.pt
Hi‑Res Fix/Upscaling: lollypop
Samplers and Schedulers: Euler a Automatic
Steps: 20 or 35
CFG: 5
Width and Height: 768×1344, 832×1216
For storytelling, specify mood and setting (e.g., “fantasy garden”, “moonlit ocean”, “urban dusk”, “cozy room”)
🩵 All examples are made in SD A1111
You can use this site to better understand the settings of the examples 🩵
I hope you enjoy! Follow for more, leave comments, put reactions and create!
You can help me with it, tips are always welcome 💌
1. This model is intended for entertainment, research, and experimental use, including NSFW content. They do not claim to copy, reproduce, or replace any existing works, characters, styles, or brands, including those protected by copyright, trademarks, or patents. Training is based on transformation and abstraction of visual features, without direct reproduction of specific protected works. All depicted characters are fictional and imagined as adults (21+); any resemblance to real people, names, or events is purely coincidental and unintentional.
2. The author is not liable for any interpretation, distribution, modification, forks, or consequences resulting from third-party use of the models. Users are solely responsible for complying with the laws of their jurisdiction and ethical standards when using the models, including the creation and sharing of NSFW content.
3. Models are provided “as is,” without any warranties — explicit or implied — regarding accuracy, safety, completeness, or fitness for a particular purpose. While the author strives for quality and reliability, they are not responsible for any indirect, incidental, or other damages that may arise from using the models.
4. Any profiles, publications, or channels not affiliated with the official PupaBoo account on Civitai are considered unauthorized. This model is shared here for research and creative purposes only. The author kindly asks users not to re-upload this model to third-party sites or platforms