- We’re introducing data2vec, the first high-performance self-supervised algorithm that learns in the same way for speech, vision and text.
- With data2vec, we’re closer to building machines that learn about different aspects of the world around them without having to rely on labeled data. Read More