RLDS
收藏资源简介:
RLDS是一个专为强化学习设计的数据集生态系统,由谷歌研究团队开发。该系统支持序列决策制定(SDM)领域的数据记录、回放、操作、注释和共享,特别适用于强化学习(RL)、从演示中学习、离线RL或模仿学习。RLDS不仅支持现有研究的再现性和新数据集的轻松生成,还加速了新研究的进展。通过提供标准和无损的数据集格式,它使得快速测试新算法在更广泛任务上的应用成为可能。此外,RLDS还提供了收集由合成代理或人类生成的数据的工具,以及检查和操作收集数据的功能。最终,通过与TFDS的集成,RLDS促进了RL数据集与研究社区的共享。
RLDS is a dataset ecosystem specifically designed for reinforcement learning, developed by the Google Research team. This system supports data logging, playback, manipulation, annotation, and sharing in the field of sequential decision-making (SDM), and is particularly suitable for reinforcement learning (RL), learning from demonstration, offline RL, and imitation learning. RLDS not only supports the reproducibility of existing research and the easy generation of new datasets, but also accelerates the advancement of new research. By providing a standard and lossless dataset format, it enables rapid testing of new algorithms across a wider range of tasks. In addition, RLDS also provides tools for collecting data generated by synthetic agents or humans, as well as functionalities for inspecting and manipulating the collected data. Ultimately, through integration with TFDS, RLDS facilitates the sharing of RL datasets with the research community.

- 1RLDS: an Ecosystem to Generate, Share and Use Datasets in Reinforcement Learning谷歌研究 · 2021年



