In this work, we demonstrate that Multi-modal Large Language Models (MLLMs) can enhance visual-language representation learning by establishing richer image-text associations for image-text datasets.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback