Residual Reinforcement Learning Environments

该数据集专为测试基于技能的强化学习方法的有效性而设计，包含了一系列机器人操作任务。该数据集涵盖了四个不同的下游任务，这些任务改编自现有环境，旨在研究并提升学习者在不同条件下掌握和适应技能的能力。这些任务涉及到使用7自由度（7-DoF）的Fetch机器人臂执行多任务操作。总的来说，这一数据集聚焦于机器人操作领域。

This dataset is specifically designed to evaluate the effectiveness of skill-based reinforcement learning approaches, and encompasses a suite of robotic manipulation tasks. It includes four distinct downstream tasks adapted from existing environments, which are intended to investigate and improve learners' capacity to acquire and adapt skills across diverse conditions. These tasks entail multi-task manipulation operations executed by a 7-degree-of-freedom (7-DoF) Fetch robotic arm. Overall, this dataset is centered on the domain of robotic manipulation.