--- license: mit tags: - pytorch - diffusers - unconditional-image-generation - diffusion-models-class --- This model is a diffusion model for unconditional image generation of [Bored Apes](https://opensea.io/collection/boredapeyachtclub). ## Example ![example.png](https://s3.amazonaws.com/moonup/production/uploads/1670089095011-630283a6b002e9a4a216f3e4.png) ## Usage ```python from diffusers import DDPMPipeline pipeline = DDPMPipeline.from_pretrained('daspartho/bored-ape-diffusion') image = pipeline().images[0] image ``` ## Training The model is trained on [this dataset](https://huggingface.co./datasets/daspartho/bored-ape) of 10000 Bored Ape images. [Notebook](https://github.com/daspartho/bored-ape-diffusion/blob/main/main.ipynb) for training the model is available on GitHub repo.