google-research/robopianist
A MuJoCo-based benchmark suite using deep reinforcement learning to train simulated anthropomorphic robot hands to play piano.
Not currently ranked — collecting fresh signals.
star history
RoboPianist is a benchmark for high-dimensional control that uses deep RL to train dual Shadow hands to play piano with precision. It provides simulation environments, RL training infrastructure, MIDI datasets, and evaluation tasks in MuJoCo. The project enables research into bimanual dexterity, contact-rich manipulation, and coordination planning.
Frequently asked
- What is google-research/robopianist?
- A MuJoCo-based benchmark suite using deep reinforcement learning to train simulated anthropomorphic robot hands to play piano.
- Is robopianist open source?
- Yes — google-research/robopianist is open source, released under the Apache-2.0 license.
- What language is robopianist written in?
- google-research/robopianist is primarily written in Python.
- How popular is robopianist?
- google-research/robopianist has 736 stars on GitHub.
- Where can I find robopianist?
- google-research/robopianist is on GitHub at https://github.com/google-research/robopianist.