s-tian/VP2

Training datasets and benchmark data corresponding to the International Conference on Learning Representations (ICLR 2023) paper "A Control-Centric Benchmark for Video Prediction".

Please see the GitHub repository at https://github.com/s-tian/vp2 for usage instructions.

This dataset includes all data initially uploaded to the Stanford Digital Repository: https://sdr.stanford.edu/works/6335 and https://sdr.stanford.edu/works/6494.