This repo contains the official implementation for the paper Gotta Go Fast When Generating Data with Score-Based Models, which shows how to generate data as fast as possible with score-based models using a well-designed SDE solver. See the blog post for more details.
This code is a heavy modification of the Generative Modeling through Stochastic Differential Equations repository.
If you want the PyTorch version, please use https://github.com/yang-song/score_sde_pytorch/blob/main/sde_lib.py along with my modified version of https://github.com/yang-song/score_sde_pytorch/blob/main/sampling.py which is here: https://github.com/AlexiaJM/score_sde_fast_sampling/blob/main/sde_sampling_torch.py.
Notes:
See the requirements. Change the settings and folders in https://github.com/AlexiaJM/score_sde_fast_sampling/blob/main/experiments.sh and run parts of the script to run the CIFAR-10, LSUN-Church, and FFHQ experiments.
The SDE solver can be found here and the loop here.
Please refer to the original code.
https://drive.google.com/drive/folders/10pQygNzF7hOOLwP3q8GiNxSnFRpArUxQ?usp=sharing
If you find the code useful for your research, please consider citing
@article{jolicoeurmartineau2021gotta,
title={Gotta Go Fast When Generating Data with Score-Based Models},
author={Alexia Jolicoeur-Martineau and Ke Li and R{\'e}mi Pich{\'e}-Taillefer and Tal Kachman and Ioannis Mitliagkas},
journal={arXiv preprint arXiv:2105.14080},
year={2021}
}
and
@inproceedings{
song2021scorebased,
title={Score-Based Generative Modeling through Stochastic Differential Equations},
author={Yang Song and Jascha Sohl-Dickstein and Diederik P Kingma and Abhishek Kumar and Stefano Ermon and Ben Poole},
booktitle={International Conference on Learning Representations},
year={2021},
url={https://openreview.net/forum?id=PxTIG12RRHS}
}
Official theme song can be found here: https://soundcloud.com/emyaze/gotta-go-fast.