TempoRL-experiment data

Learning curve data, as well as commands to generate the data and final trained policies of all experiments of the TempoRL paper.<br><b>Readme.md</b> -&gt; Contains description of the folder structures in all three archivesexperiments.tar.gz -&gt; Archive containing all experiment results with the following folders:<br><b> atari</b> -&gt; Contains all results on Atari games<b> featurized_results</b> -&gt; Contains results on LunarLander-v2 and MountainCar-v0<b> tabular_results</b> -&gt; Contains results for the tabular experiments.<br>DDPG results have been added in a separate zip<br><b>tempoCar1.mp4</b> and <b>tempoCar2.mp4</b> show learned policies of tempoRL in the mountain car environment. The agent blinks bright blue when it decides to make a new decision.<b>tempoQBert.mp4</b> shows a tempoRL policy in action on QBert.<br><br>

本数据集涵盖TempoRL论文所有实验的学习曲线数据、数据生成命令以及最终训练完成的智能体策略。
<b>Readme.md</b>：用于说明三个归档文件的文件夹结构。
<b>experiments.tar.gz</b>：包含全部实验结果，内含以下子文件夹：
<b>atari</b>：收录雅达利（Atari）游戏实验的全部结果；
<b>featurized_results</b>：收录月球着陆器-v2（LunarLander-v2）与山地车-v0（MountainCar-v0）实验的结果；
<b>tabular_results</b>：收录表格型强化学习实验的结果。
深度确定性策略梯度（Deep Deterministic Policy Gradient, DDPG）的实验结果已单独打包为压缩文件。
<b>tempoCar1.mp4</b>与<b>tempoCar2.mp4</b>：展示TempoRL在山地车环境中学到的运行策略，智能体在决定进行新决策时会闪烁亮蓝色。
<b>tempoQBert.mp4</b>：展示TempoRL策略在QBert游戏中的实际运行效果。