eloialonso/diamond
DIAMOND is a reinforcement learning agent trained entirely within a diffusion-based world model, demonstrated on Atari games and CSGO.
Not currently ranked — collecting fresh signals.
star history
DIAMOND (DIffusion As a Model Of eNvironment Dreams) trains a reinforcement learning agent to act within an imagined environment generated by a diffusion model. The world model learns to predict future frames from current observations and actions, enabling the agent to plan and learn via simulated experience rather than real environment interactions. The project provides pretrained Atari and CSGO world models along with training code for both the diffusion model and the RL policy.
Frequently asked
- What is eloialonso/diamond?
- DIAMOND is a reinforcement learning agent trained entirely within a diffusion-based world model, demonstrated on Atari games and CSGO.
- Is diamond open source?
- Yes — eloialonso/diamond is open source, released under the MIT license.
- What language is diamond written in?
- eloialonso/diamond is primarily written in Python.
- How popular is diamond?
- eloialonso/diamond has 2.1k stars on GitHub.
- Where can I find diamond?
- eloialonso/diamond is on GitHub at https://github.com/eloialonso/diamond.