Generative Transformers

BERT (Bidirectional Encoder Representations from Transformers)

Bidirectional pre-trained Transformer model with Masked Language Modeling to understand context in both directions of a sequence.

← Tillbaka