Ddpm Church 256

🚀 Denoising Diffusion Probabilistic Models (DDPM)

This project presents high - quality image synthesis results using diffusion probabilistic models, offering a new approach for unconditional image generation.

🚀 Quick Start

Inference

DDPM models can utilize discrete noise schedulers like:

• scheduling_ddpm
• scheduling_ddim
• scheduling_pndm

It's important to note that while the ddpm scheduler produces the highest - quality results, it also has the longest inference time. For a better balance between quality and inference speed, you might consider using the ddim or pndm schedulers.

💻 Usage Examples

Basic Usage

# !pip install diffusers
from diffusers import DDPMPipeline, DDIMPipeline, PNDMPipeline

model_id = "google/ddpm-church-256"

# load model and scheduler
ddpm = DDPMPipeline.from_pretrained(model_id)  # you can replace DDPMPipeline with DDIMPipeline or PNDMPipeline for faster inference

# run pipeline in inference (sample random noise and denoise)
image = ddpm().images[0]

# save image
image.save("ddpm_generated_image.png")

For more detailed information, please refer to the official inference example

Training

If you wish to train your own model, please check out the official training example

📚 Documentation

Paper and Authors

• Paper: Denoising Diffusion Probabilistic Models
• Authors: Jonathan Ho, Ajay Jain, Pieter Abbeel

Abstract

We present high quality image synthesis results using diffusion probabilistic models, a class of latent variable models inspired by considerations from nonequilibrium thermodynamics. Our best results are obtained by training on a weighted variational bound designed according to a novel connection between diffusion probabilistic models and denoising score matching with Langevin dynamics, and our models naturally admit a progressive lossy decompression scheme that can be interpreted as a generalization of autoregressive decoding. On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state - of - the - art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to ProgressiveGAN.

📦 Samples

📄 License

This project is licensed under the Apache 2.0 license.