Gym-MuJoCo datasets for offline reinforcement learning
收藏科学数据银行2025-07-02 更新2026-04-23 收录
下载链接：
https://www.scidb.cn/detail?dataSetId=0e9ae388c88845298a34f9133f2aa71d
下载链接
链接失效反馈官方服务：
资源简介：
The Gym-mujoco dataset originates from the open-source D4RL bechmark and comprises three tasks: Hopper, Halfcheetahh, and Walker2d. Each task includes five different levels of datasets: random, medium, medium-replay, medium-expert, and expert. The provided Gym-mujoco dataset is in hdf5 format for easy data viewing and usage. You can also download by installing D4RL environment, reference site at: https://github.com/Farama-Foundation/D4RL?tab=readme-ov-file
提供机构：
Xuesong Wang; Huanting Feng; Yuhu Cheng
创建时间：
2025-07-02
