RoboPianist
Introduced by Zakka et al. in RoboPianist: Dexterous Piano Playing with Deep Reinforcement Learning
RoboPianist is a benchmarking suite for high-dimensional control, targeted at testing high spatial and temporal precision, coordination, and planning, all with an underactuated system frequently making-and-breaking contacts. The proposed challenge is mastering the piano through bi-manual dexterity, using a pair of simulated anthropomorphic robot hands. The initial version covers a broad set of 150 variable-difficulty songs.
Source: RoboPianist: A Benchmark for High-Dimensional Robot ControlPapers
| Paper | Code | Results | Date | Stars |
|---|
Dataset Loaders
No data loaders found. You can
submit your data loader here.
No data loaders found. You can
submit your data loader here.
