Awesome-World-Models

A curated survey and resource index for World Model research. 890+ GitHub stars.

Awesome-World-Models is a research-oriented survey and resource index (Survey & Knowledge Base) for the World Model research direction.

Overview

This project systematically curates representative works covering:

  • Video prediction and generation
  • Action-conditioned dynamics modeling
  • Embodied AI applications
  • Reinforcement learning with world models

The project provides a structured research entry point for the World Model and Embodied AI community, and is widely used for literature surveys, method comparisons, and research direction planning.

My Contributions

  • Participated in maintaining and continuously expanding World Model related research resources
  • Systematically organized representative works across multiple research directions