Open-source TinyWorlds replicates DeepMind's Genie world model, hits 1.3k GitHub stars
tom_doerr · x · 2026-08-11
Developer AlmondGod released TinyWorlds, a minimal autoregressive world model implementation based on DeepMind's Genie architecture. The project has gained 1.3k stars and 102 forks on GitHub.
Genie's key innovation is learning from action-less internet video by inferring actions between frames, enabling scalable world models. TinyWorlds aims to help developers understand this autoregressive unsupervised method, featuring video tokenizer, action tokenizer, and dynamics model components.
The project provides complete code, configs, and training scripts with WandB integration, suitable for researching world models and video generation.
More from Research
- After Math Falls: AI Researchers Face the Question of What Comes Next — ziv_ravid · 2026-08-12
- HALO: Human-AI collaborative system for drug discovery molecular hypothesis generation — _xiang_chen_ · 2026-08-12
- Solving RAG Bottlenecks: A 2026 Open-Source Guide to PDF Table Parsing — AvenueJay · 2026-08-12
- AI pushes mathematical bound for packing 17 squares to 4.456575 — stanislavfort · 2026-08-12
- Founder Uses ChatGPT to Design mRNA Cancer Vaccine for His Dog, Launches YC-Backed Gamgee — ycombinator · 2026-08-12
- CSCW 2026 Workshop to Explore AI's Impact on Open Source — manoelribeiro · 2026-08-12