Aniimage-1

Aniimage-1 is the first latent diffusion model developed by 8BitStudio. The model is a 256x256 anime image generation model trained from scratch using a UNet + VAE + CLIP architecture at 449.1 million parameters. Aniimage-1 has been trained on 830,001 anime images from Danbooru. It is not based on any existing models, the UNet is trained from scratch.

Model Details

Resolution	256×256
Architecture	Latent Diffusion (UNet + VAE + CLIP)
Parameters	441.9M
Training Steps	88,000
Batch Size	64
Dataset	~830K curated anime images from Danbooru
GPU	NVIDIA RTX 5060 Ti 16GB
Scheduler	DDIM or DPM ++ 2M

Requirements

• GPU: ~3.4 GB VRAM minimum (recommend 4+ GB)
• CPU: ~2 GB RAM. Image generation is extremely slow on CPU.

Quick Start

after downloading, install the dependencies.

pip install torch torchvision diffusers transformers safetensors pillow huggingface_hub
python generate_hf.py

recommended settings: Scheduler on DPM++ 2M with 25 steps and a CFG of 7.5. If you are using Euler, 30 steps and 5 CFG works the best. recommended negative prompt: "low quality, ugly, blurry, distorted, deformed, bad anatomy, bad proportions, extra limbs, missing limbs, watermark, text, signature, washed out, flat colors, manga panel, disfigured, poorly drawn, jpeg artifacts, cropped, out of frame"

Prompting

Aniimage uses plain text captions as well as Danbooru tags.

An example of acceptable captions:

"A smiling anime girl with red hair and a school uniform"

"1girl, solo, smile, red_hair, school_uniform, anime_coloring"

Long prompts and enhanced prompts usually result in worse overall results. Try to keep your prompts short.

Capabilities

• Anime character generation with varied hair colors and styles
• School uniforms, fantasy outfits, maid dresses, and more
• Background scenes: cherry blossoms, night sky, interiors, nature

Limitations

• 256×256 resolution — fine details like hands and small features can be rough
• Faces can sometimes look similar or 'melty' across different prompts
• Little to none NSFW content — trained on mostly SFW dataset only
• Does worse when generating men due to dataset bias

Oddities

_{Prompt: una chica anime con cabello largo rosa
(an anime girl with long pink hair)}

As shown above, Aniimage-1 can sometimes understand Spanish prompts. Longer prompts may be less reliable, but Spanish support does work in some cases. Strangely, Japanese does not work at all, producing images that look extremely distorted and not conforming to the prompt at all.

_{Prompt: anime girl with short silver hair sitting by a window}

Aniimage-1 has a bad habit of hiding hands. This is likely because hands turn out melty, if you don't specify it, hands will likely be concealed, hidden, or simply no hands at all.

_{Prompt: 1girl, solo, school_uniform, sailor_collar, pleated_skirt, brown_hair, smile, classroom}

When using Danbooru tags, it is much more likely to have black boxes or bars around the sides of an image as seen above. This may be caused by how the model’s training data was filtered, possibly from not cropping images that were smaller than or outside its expected aspect ratio.

_{Prompt: A portrait of a solo anime girl with brown hair wearing a school uniform with a sailor collar and a pleated skirt in a classroom}

When using plain English text, you are much less likely to experience this issue.

What's Next

Aniimage-2 is complete!

It features a native resolution of 512×512 and was trained on 1.2 million unique images, compared with the 830,000 images used to train Aniimage-1.

Aniimage-2 also switches from Aniimage-1’s epsilon-prediction objective to v-prediction and uses a zero-terminal signal-to-noise ratio. It also uses uniform timestep sampling, while the upcoming Aniimage-3 will adopt logit-normal noise-level sampling as part of its flow-matching objective.

License

Apache 2.0