BurakCanBiner/SonicDiffusion

SonicDiffusion: Audio-Driven Image Generation and Editing with Pretrained Diffusion Models

Official Implementation of SonicDiffusion: Audio-Driven Image Generation and Editing with Pretrained Diffusion Models.

Updates

03.05.2024: Our code and demo are released.

Demo

Getting Started

Prerequisites

$ conda create -n "myenv" python=3.11
$ git clone git@github.com:BurakCanBiner/SonicDiffusion.git
$ cd SonicDiffusion
$ pip install -r requirements.txt

Pretrained Models

Pretrained models are available at the following link.

https://huggingface.co/spaces/burakcanbiner/SonicDiffusion/tree/main/ckpts

By default, we assume that all models are downloaded and saved to the directory ckpts.

Citation

@misc{biner2024sonicdiffusion,
      title={SonicDiffusion: Audio-Driven Image Generation and Editing with Pretrained Diffusion Models}, 
      author={Burak Can Biner and Farrin Marouf Sofian and Umur Berkay Karakaş and Duygu Ceylan and Erkut Erdem and Aykut Erdem},
      year={2024},
      eprint={2405.00878},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}