Last updated 07-30-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Waifu-diffusion on Hugging Face
Waifu-diffusion on Hugging Face turns text prompts into anime-style images through a Stable Diffusion model fine-tuned on high-quality anime artwork. You load hakurei/waifu-diffusion with the Diffusers library in Python, pass a prompt, and get a generated image back. The model card includes ready-to-run code for CUDA, Apple MPS, Google Colab, and local apps like Draw Things and DiffusionBee.
Compared with general-purpose Stable Diffusion checkpoints, this model is tuned specifically for anime aesthetics rather than photorealism. That narrow focus makes it a common base for fan art, character concepts, and stylized illustrations, while general SD models often need heavier prompting to stay on-anime. The v1.4 weights ship under CreativeML OpenRAIL-M with 2,500+ community likes and about 1,166 downloads in the last month on Hugging Face.
The model weighs 0.9 billion parameters in Safetensors format and works with the StableDiffusionPipeline in Diffusers. Creators can fine-tune or adapt it further, with four adapter models, two finetunes, and one quantization listed in the Hugging Face model tree.
Waifu-diffusion suits illustrators, hobbyists, and developers who want anime image generation they can run locally or in notebooks without building a model from scratch.
0.9B parameter latent text-to-image model fine-tuned on anime artwork
Runs with Diffusers StableDiffusionPipeline and example CUDA code on the model card
2,500+ likes and 1,166 downloads last month on Hugging Face
CreativeML OpenRAIL-M license allows commercial redistribution with use restrictions
Google Colab and Kaggle notebook links included for cloud testing
Imports into Draw Things and DiffusionBee local apps from the model page
100+ Hugging Face Spaces build on hakurei/waifu-diffusion as a base model
Free open weights with a large anime-focused training bias
Works with standard Diffusers pipelines and published example code
Colab, Kaggle, and local app import links lower the setup barrier
Strong community adoption with 2,500+ likes on Hugging Face
Adapter and finetune ecosystem extends the base model
Requires your own GPU or cloud notebook for practical inference
Anime-specific tuning produces poor results for photorealistic prompts
OpenRAIL-M license restricts deliberate harmful or illegal outputs
Not deployed on Hugging Face Inference Providers by default
What does the hakurei/waifu-diffusion model do?
Waifu-diffusion on Hugging Face is hakurei/waifu-diffusion, a latent text-to-image model fine-tuned on anime images. You load it with the Diffusers library and generate images from text prompts on your own GPU or in Colab.
Does waifu-diffusion cost money to download?
No. Waifu-diffusion on Hugging Face is free to download and use under the CreativeML OpenRAIL-M license. Hugging Face hosts the weights at no charge, though running inference requires your own compute.
How do I install waifu-diffusion locally?
Waifu-diffusion on Hugging Face runs with pip install diffusers transformers accelerate, then StableDiffusionPipeline.from_pretrained('hakurei/waifu-diffusion'). The model card includes Python examples for CUDA and links to Colab and local apps.
What license covers waifu-diffusion weights?
Waifu-diffusion on Hugging Face uses the CreativeML OpenRAIL-M license. You cannot use it to deliberately produce illegal or harmful content, but you may use outputs freely and redistribute weights commercially with the same restrictions.
Can businesses use waifu-diffusion outputs?
Waifu-diffusion on Hugging Face allows commercial use and redistribution under CreativeML OpenRAIL-M. You must pass the same use restrictions to downstream users and include a copy of the license when redistributing the weights.
What GPU do I need for waifu-diffusion?
Waifu-diffusion on Hugging Face runs on NVIDIA CUDA GPUs in the published examples, with an MPS option noted for Apple devices. You can also test it in Google Colab or Kaggle notebooks linked from the model page.