Isaac Gym Benchmark Tasks

该数据集为强化学习提供了一系列基准任务，包括蚂蚁、人形、ANYmal、Shadow Hand、Allegro Hand以及Franka立方体堆叠等。此外，还包含了两个附加任务：一个是基于视觉的球平衡任务，另一个是使用DClaw Hand的富含接触的灵巧操作任务。该数据集规模宏大，拥有数万个并行环境，旨在推动强化学习任务的发展。

This dataset provides a suite of benchmark tasks for reinforcement learning, including Ant, Humanoid, ANYmal, Shadow Hand, Allegro Hand, and Franka Cube Stacking, among others. Additionally, it includes two supplementary tasks: one is a vision-based ball balancing task, and the other is a contact-rich dexterous manipulation task using the DClaw Hand. This dataset has a large scale with tens of thousands of parallel environments, aiming to advance the development of reinforcement learning tasks.