FlagOpen/Robo-Dopamine
Process reward modeling trained on 35M dataset
View on index · View in 3D Map
Related repos
- RoboBrain 2.0: open-source model for robotic reasoning, planning, spatial and temporal perception, and scene understanding
- Trains robotic reward models using trajectory comparisons
- Cognitive memory engine for agents
- Hierarchical embodied OS with shared memory for multi-robot systems
- Infrastructure for agentic reinforcement learning
- Platform for robot learning across simulators and embodiments
- open datasets and models for training reasoning AI
- Generates endless robot training tasks and environments
- Reinforcement learning tuning for reasoning LLM agents
- Robotics models pre-trained on 10k hours of data
