Multimodal QA

Modality-to-Modality Alignment

Learning process that matches segments of one modality (e.g., a sentence) with relevant segments of another (e.g., an image region).

← Zurück