Synth: Boosting Visual-Language Models with Synthetic Captions and Image Embeddings

Arxiv Papers - Synth: Boosting Visual-Language Models with Synthetic Captions and Image Embeddings