Demonstrations for Behavior Cloning
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/GT-STAR-Lab/constrained-rl-dexterous-manipulation
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了25个通过CyberGlove III收集的演示数据，这些数据被用于预训练强化学习策略。此外，该数据集在应用约束强化学习算法之前，也被用来对策略网络进行预训练。所涉及的任务是针对强化学习的行为克隆。
This dataset consists of 25 sets of demonstration data collected using CyberGlove III, which is utilized for pre-training reinforcement learning policies. Furthermore, this dataset is also employed to pre-train policy networks prior to the application of constrained reinforcement learning algorithms. The task involved herein is behavior cloning for reinforcement learning.
提供机构：
GT-STAR Lab
