Palette: Image-to-Image Diffusion Models

🤖 Plain-English Summary

This paper develops a unified framework for image-to-image translation based on conditional diffusion models and evaluates this framework on four challenging image-to-image translation tasks, namely colorization, inpainting, uncropping, and JPEG restoration. Finally, we show that a generalist, multi-task diffusion model performs as well or better than task-specific specialist counterparts.

🔑 Key Findings

• Our simple implementation of image-to-image diffusion models outperforms strong GAN and regression baselines on all tasks, without task-specific hyper-parameter tuning, architecture customization, or any auxiliary loss or sophisticated new techniques needed.
• We uncover the impact of an L2 vs.
• L1 loss in the denoising diffusion objective on sample diversity, and demonstrate the importance of self-attention in the neural architecture through empirical studies.

🔗 Related Resources