4 ms·Large-Scale Self-Supervised Pre-Training Across Tasks, Languages, and Modalities3 points by axiomdata316 5y ago