The author presents a toolkit, Imitation Learning Datasets, to address challenges in creating datasets for training agents and benchmarking imitation learning techniques.