Rethinking Video-Text Understanding Retrieval from Counterfactually Augmented Data - View it on GitHub