aditijc/snooker-testbed-canary-7049627-stage0-v1

该数据集与使用PPO（近端策略优化）算法进行的斯诺克实验台实验相关。数据集包含10行和9列，包括步骤、课程阶段、平均得分、最高得分、平均击球次数、平均效率、平均犯规率、回合数和源文件等详细信息。实验涉及一个课程引导，分为0到4阶段，其中阶段0将1个红球放置在球杆和右上角袋之间的中间位置。该数据集是Phase-3 Option-B的金丝雀测试的一部分，实验未通过go/no-go标准，但与之前的金丝雀相比，阶段0和1的进展更快。数据集使用stable-baselines3 PPO生成，具有特定的超参数，并附有详细的实验参数和结果文档。

This dataset is related to a snooker testbed experiment using the PPO (Proximal Policy Optimization) algorithm. The dataset contains 10 rows and 9 columns, including details such as step, curriculum_stage, mean_score, max_score, mean_shots, mean_efficiency, mean_foul_rate, episodes, and source_file. The experiment involves a curriculum bootstrap with stages 0 to 4, where stage 0 places 1 red midway between the cue and top_right pocket. The dataset is part of a canary test for Phase-3 Option-B, and the experiment failed the go/no-go criteria but showed faster progression through stages 0 and 1 compared to a previous canary. The dataset is generated using stable-baselines3 PPO with specific hyperparameters and is documented with detailed experiment parameters and results.