dm_control: Software and Tasks for Continuous Control
This provides a standardized toolkit for researchers in reinforcement learning to develop and benchmark agents in continuous control tasks, though it is incremental as it builds on existing simulation frameworks.
The authors introduced dm_control, a software package with Python libraries and task suites for reinforcement learning in articulated-body simulation, providing tools like a MuJoCo wrapper, PyMJCF, Composer, Control Suite, Locomotion framework, and manipulation tasks, which is publicly available on GitHub.
The dm_control software package is a collection of Python libraries and task suites for reinforcement learning agents in an articulated-body simulation. A MuJoCo wrapper provides convenient bindings to functions and data structures. The PyMJCF and Composer libraries enable procedural model manipulation and task authoring. The Control Suite is a fixed set of tasks with standardised structure, intended to serve as performance benchmarks. The Locomotion framework provides high-level abstractions and examples of locomotion tasks. A set of configurable manipulation tasks with a robot arm and snap-together bricks is also included. dm_control is publicly available at https://www.github.com/deepmind/dm_control