Ultimately, VideoPoet has become capable of predicting video tokens similar to a traditional LLM model would predict text tokens. In the aforementioned paper, Google’s researchers defined that VideoPoet differs from […]