thu-ml/RoboticsDiffusionTransformer
A 1-billion parameter diffusion transformer trained on 1M+ robot episodes that predicts bimanual manipulation actions from language instructions and RGB images.
Not currently ranked — collecting fresh signals.
star history
RDT-1B is a foundation model for robot manipulation that uses a diffusion transformer architecture to generate sequences of robot actions. Given natural language instructions and multi-view RGB observations, the model predicts next actions for dual-arm robotic systems. It is pre-trained on over 1 million multi-robot episodes and can be fine-tuned for specific bimanual tasks. The implementation includes PyTorch model code, training scripts with DeepSpeed, pre-trained checkpoints on HuggingFace, and real-robot deployment examples.
Frequently asked
- What is thu-ml/RoboticsDiffusionTransformer?
- A 1-billion parameter diffusion transformer trained on 1M+ robot episodes that predicts bimanual manipulation actions from language instructions and RGB images.
- Is RoboticsDiffusionTransformer open source?
- Yes — thu-ml/RoboticsDiffusionTransformer is open source, released under the MIT license.
- What language is RoboticsDiffusionTransformer written in?
- thu-ml/RoboticsDiffusionTransformer is primarily written in Python.
- How popular is RoboticsDiffusionTransformer?
- thu-ml/RoboticsDiffusionTransformer has 1.7k stars on GitHub.
- Where can I find RoboticsDiffusionTransformer?
- thu-ml/RoboticsDiffusionTransformer is on GitHub at https://github.com/thu-ml/RoboticsDiffusionTransformer.