Atari and OpenAI Gym

该数据集包含了在Atari和OpenAI Gym环境中进行的实验，这些实验采用了DQN（离策略）和PPO（在策略）算法。该数据集用于评估天机系统相较于现有DRL训练系统（如ApeX、RLlib和XingTian）的性能。它涵盖了多个环境中的游戏，任务是对强化学习进行训练和评估。

This dataset contains experiments conducted in Atari and OpenAI Gym environments, using DQN (off-policy) and PPO (on-policy) algorithms. It is employed to evaluate the performance of the Tianji system relative to existing deep reinforcement learning (DRL) training systems including ApeX, RLlib, and XingTian. The dataset covers games across multiple environments, with tasks designed for training and evaluating reinforcement learning models.