Expert Collision-Free Data

该数据集包含了用于训练模仿学习模型、模仿专家行为的无碰撞专家数据。此外，该数据集还在受控环境中用于评估各种强化学习和模仿学习方法的表现。该数据集规模宏大，共收集了163,000个数据点用于训练，其任务专注于自动驾驶赛车领域的模仿学习。

This dataset provides collision-free expert data for training imitation learning models to replicate expert behaviors. It is also used in controlled environments to evaluate the performance of various reinforcement learning and imitation learning methods. Boasting a large scale, this dataset has collected a total of 163,000 training data points, with its tasks focusing on imitation learning in the autonomous racing domain.