hojonathanho/diffusion

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

5300 stars · 492 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

The official reference implementation of Denoising Diffusion Probabilistic Models (DDPM) from the 2020 paper by Jonathan Ho et al., written in TensorFlow 1.15. It provides training and evaluation scripts for diffusion-based image generation on TPUs, with pretrained models and samples available for download.

Use cases

• reproduce the DDPM paper results
• train a denoising diffusion model on image datasets
• generate images with a pretrained diffusion model
• study the original DDPM codebase for research
• evaluate diffusion models with FID and other metrics

When to choose

• you need the canonical DDPM implementation for research reproduction
• you have TPU/GCP infrastructure and legacy TensorFlow 1.x environments
• you want to study the original diffusion model code

When to avoid

• you want modern PyTorch or JAX diffusion libraries like Hugging Face diffusers
• you need a maintained, licensed, production-ready image generation tool
• you cannot run TensorFlow 1.15 and Python 3.5

Facets

library · maturity maintenance

machine-learning deep-learning image-processing deep-learning machine-learning image-processing python diffusion-models ddpm generative-models tensorflow research-code image-generation gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="hojonathanho/diffusion")