rllab benchmark
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/roosephu/slbo
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一组基于使用MuJoCo物理模拟器的环境的连续控制任务集合。它包含了从特定区间均匀采样速度的任务。该数据集被用于评估所提出算法与MAML算法的性能对比。任务规模涵盖了各种标准的MuJoCo模型，包括半猎豹、蚂蚁和简易人形。具体任务包括目标速度任务以及前进/后退任务。
This dataset comprises a set of continuous control tasks built on environments utilizing the MuJoCo physics simulator. All tasks within the dataset have their velocities uniformly sampled from specified intervals. This dataset is designed to enable performance comparison between the proposed algorithm and the MAML algorithm. The task suite covers a range of standard MuJoCo robotic models, including HalfCheetah, Ant, and Simple Humanoid. Specific task types include target velocity tracking tasks and forward/backward locomotion tasks.
