Abstract page for arXiv paper 2212.05051: VindLU: A Recipe for Effective Video-and-Language Pretraining