Diffusion Models

• Diffusion models: https://github.com/lucidrains/denoising-diffusion-pytorch
  • implement the classifier-free guidance diffusion
  • denoising diffusion models
• https://github.com/acids-ircam/diffusion_models
  • interactively show the score matching process and Langevin dynamics (very nice Jupyter notebook illustration)
  • implement the DDIM which accelerates the reference process
• https://github.com/openai/improved-diffusion
  • Many recent advanced diffusion models are modified from this repo.
• Classifier-guided diffusion model: https://github.com/openai/guided-diffusion/tree/22e0df8183507e13a7813f8d38d51b072ca1e67c

Neural Compression

• How to compute bit per pixel (bpp)?

  • neural_compression/neural_compression_example.ipynb

• https://huzi96.github.io/compression-bench.html

  • A collection of neural compression methods
  • Includes the Kodak dataset

GAN

• GAN Zoo: https://github.com/eriklindernoren/PyTorch-GAN

VAE

• VAE Zoo: https://github.com/AntixK/PyTorch-VAE