Generative Self-Supervision

Generative Pre-training

Self-supervised approach where the model learns by generating or reconstructing data, developing a deep understanding of the underlying distribution.

← Terug