diff-usion/awesome-diffusion-models

★ 12,321HTMLAudience · researcherComplexity · 1/5LicenseSetup · easy

Mindmap

mindmap
  root((Awesome Diffusion))
    What it does
      Paper index
      Tutorial links
      Video lectures
      Notebook demos
    Application Areas
      Image generation
      Audio synthesis
      Natural language
      Molecular design
    Use Cases
      Research discovery
      Field overview
      Beginner learning
    Audience
      Researchers
      ML practitioners

mindmap root((Awesome Diffusion)) What it does Paper index Tutorial links Video lectures Notebook demos Application Areas Image generation Audio synthesis Natural language Molecular design Use Cases Research discovery Field overview Beginner learning Audience Researchers ML practitioners

Click or tap to explore — scroll the page freely

Things people build with this

USE CASE 1

Find research papers on a specific diffusion model application like medical imaging, text-to-speech, or molecular design.

USE CASE 2

Follow a learning path from beginner blog posts to recorded university lectures without assuming graduate-level math.

USE CASE 3

Run diffusion model experiments in a browser using linked Jupyter notebooks, no local setup required.

USE CASE 4

Track the state of diffusion model research across image, audio, language, and scientific domains in one place.

Getting it running

Difficulty · easy Time to first run · 5min

Use freely for any purpose including commercial use, as long as you keep the copyright notice.

In plain English

Diffusion models are a class of AI that generate new images, audio, and other content by learning how to gradually add and remove random noise from data. This repository is a curated index of research papers, blog posts, videos, lectures, and runnable tutorials on the topic, organized into sections so that someone can find resources matching their background and area of interest. The collection spans a wide range of applications. On the vision side, papers cover image generation, classification, segmentation, image translation, medical imaging, 3D vision, and adversarial robustness. For audio, the list includes work on generation, voice conversion, speech enhancement, sound separation, and text-to-speech synthesis. There are also dedicated sections on natural language, time series forecasting, graph generation, molecular design, and reinforcement learning, showing how broadly the underlying technique has been applied across different fields. Beyond technical papers, the repository links to introductory blog posts written to make the core ideas accessible without assuming a graduate-level math background. Companion resources include YouTube videos, recorded university lectures, and Jupyter notebooks that let someone run diffusion experiments directly in the browser. A separate companion website hosts a version of the list that may be more current than what appears on the GitHub page. The list is maintained under an MIT license. It is not a software package and contains no runnable code of its own. Its purpose is to serve as a reference index for anyone who wants to understand where diffusion model research stands, find a starting point for learning, or locate a specific paper on a particular sub-topic. The full README is longer than what was shown.

Copy-paste prompts

Prompt 1

I want to learn about diffusion models from scratch. Based on the awesome-diffusion-models list, recommend the best beginner blog posts and introductory videos to start with.

Prompt 2

Find me papers from awesome-diffusion-models that apply diffusion models to medical image segmentation or analysis.

Prompt 3

Which Jupyter notebooks from the awesome-diffusion-models list can I run in a browser without a GPU to try diffusion experiments hands-on?

Prompt 4

Summarize the key conceptual differences between DDPM-style diffusion models and score-based generative models using resources from this list.

Prompt 5

What papers on the awesome-diffusion-models list cover audio generation, specifically voice conversion and text-to-speech synthesis?

Open on GitHub → Explain another repo

← diff-usion on gitmyhub — every repo by this author, as a profile.

Verify against the repo before relying on details.