Multi-Modal Transformers

Vision-Language Transformer

Transformer architecture specifically designed to jointly understand and generate visual and textual content, using shared or separate encoders for each modality.

← Tillbaka