Skill Options Dataset

该数据集是由LLM生成的拟议机器人任务集，用于在模拟环境中评估机器人臂的技能获取能力。它包含了获取技能的各项度量，包括成功评估和错误评估的技能候选者，并报告了学习过程。该数据集规模较大，涵盖了通过强化学习提出和评估的多项任务，旨在研究机器人操作中的任务提议与技能获取。

This dataset is a collection of proposed robotic tasks generated by large language models (LLMs), designed to evaluate the skill acquisition capabilities of robotic arms in simulated environments. It encompasses various metrics for skill acquisition, including both successfully and erroneously evaluated skill candidates, and documents the learning process. With a large scale, this dataset covers multiple tasks proposed and evaluated via reinforcement learning, aiming to investigate task proposal and skill acquisition in robotic manipulation.