BERT et ses Variantes

VideoBERT

Multimodal extension of BERT learning joint video-text representations. Performs pre-training on visual and linguistic tokens for video understanding.

← Tillbaka