Bofuri Full

🚀 Bofuri Full Model

This is a text - to - image model trained on 33 different concepts from the anime Bofuri: I Don't Want to Get Hurt, so I'll Max Out My Defense, capable of generating high - quality anime - style images.

🚀 Quick Start

This model is shared in both diffuser and safetensors format. Intermediate checkpoints are also shared in ckpt format in the directory checkpoints.

✨ Features

• Concept - based Generation: Trained on 33 different concepts from the anime, allowing for diverse image generations.
• Anime Style: Capable of generating high - quality anime - style images, suitable for creating 4K wallpapers and official - art - like images.
• Multi - concept Support: Although there may be some blending issues when using multiple concepts, it can still generate images with multiple characters or elements.

💻 Usage Examples

Basic Usage

Here are some example generations with different prompts:

Prompt: BoMaple uniform BoSally unfirom, yuri, in classroom, 4K wallpaper, beautiful eyes

Prompt: 2girls, BoMay BoYui, yuri, half body, floating in the sky, cloud, sparkling eyes, 4K wallpaer, anime coloring, official art

Prompt: BoKanade casting magic, 4K wallpaper, outdoors

Advanced Usage

• Negative Prompt: The negative prompt is mostly variations of: bad hands, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry
• Prompt Format: During training, the concept names are put at the beginning of the images separated only by spaces, but not doing so seems to work as well. Putting aniscreen after the concept names would reinforce the anime style.

Prompt: (BoMaple black armor) BoSally turtleneck BoKasumi, 3girls, 4K wallpaper, ahoge, black hair, brown hair, outdoors, long hair

📚 Documentation

Concepts

The 33 concepts are listed in concept_list and demonstrated below.

BoMaple + BoSally + The following use the full name of the concepts

Tips for Specific Concepts:

• Expect bad results for BoMaple sheep form and non - human concepts.
• For BoKasumi sarashi, adding bandages seems to help.
• For BoMaple pajama, you can add stripe for more similarity to the pajama appearing in anime.

More Generations

Here are some more generation examples:

Prompt: BoMaple black armors aniscreen, 1girl solo, Hydra in the sky, light purple eyes, 4K wallpaper

Prompt: BoMaple black armors near small turtle syrup, sitting with knees up on rock looking at viewer, turtle shell, beautiful hand in glove, in front of trees , outdoors, close - up, 4K wallpaper

Prompt: BoMaple pajama stripe, sitting on bed with barefoot, in girl's room, detailed and fancy background, sparkling purple eyes, hand on bed, 4K wallpaper

Prompt: BoFrederica, cowboy shot, in rubble ruins, ((under blue sky)), cinematic angle, dynamic pose, oblique angle, 4K wallpaer, anime coloring, official art

Prompt: Turtle Syrup Fox Oboro next to each other simple background white background, animals

Failures:

Model Merging: You can always get different styles via model merging.

Dataset Description

The dataset is prepared via the workflow detailed here: https://github.com/cyber - meow/anime_screenshot_pipeline

It contains 27031 images with the following composition:

• 7752 bofuri images mainly composed of screenshots from the first season and of the first three episods of the second season.
• 19279 regularization images which intend to be as various as possible while being in anime style (i.e. no photorealistic image is used).

Note that the model is trained with a specific weighting scheme to balance between different concepts so that every image does not weight equally. After applying the per - image repeat, we get around 200K images per epoch.

Training

Training is done with EveryDream2 trainer with ACertainty as base model. The following configuration is used:

• Resolution: 512
• Learning Rate Scheduler: Cosine learning rate scheduler, lr 2.5e - 6
• Batch Size: 4
• Conditional Dropout: 0.05
• Beta Scheduler: Change beta scheduler from scaler_linear to linear in config.json of the scheduler of the model
• Clip Skip: 1

The released model is trained for 57751 steps, but among the provided checkpoints, all the three starting from 34172 steps seem to work reasonably well.

🔧 Technical Details

The model is trained using a specific weighting scheme to balance different concepts from the anime. The training process uses the EveryDream2 trainer with ACertainty as the base model, and specific configurations are set for resolution, learning rate scheduler, batch size, etc.

📄 License

This model is released under the creativeml - openrail - m license.