Diffusion_PathTracing_Denoiser

Creating a diffusion model from scratch in PyTorch and training it for denoising Path Tracing renderings.

TODO List

• Implement the diffusion model in PyTorch.
• Implement the training loop.
• Overfit on the Cornell Box dataset generated using mitusba.
• Improve speed using https://github.com/hutaiHang/Faster-Diffusion
• https://arxiv.org/html/2307.03992v3/#S5 Diffusion for General Denoising

References

• https://medium.com/@mickael.boillaud/denoising-diffusion-model-from-scratch-using-pytorch-658805d293b4
• https://lilianweng.github.io/posts/2021-07-11-diffusion-models/#reverse-diffusion-process
• https://theaisummer.com/diffusion-models/#architecture