Can I get the image-text fusion feature embeddings in this model only by a image?

[Linn0910]

Hello! Thanks for your work! Since llava has a generation ability, so I want to konw can this model get the image-text fusion feature embeddings in this model only by a image.
Thanks for your time and help!
Best regards!

[kongds]

Thank you for your interest in our work.

Did you mean embedding text from an image, such as OCR?
We have explored similar approaches in our paper, such as rendering captions onto images in Section 4.2. However, we have not attempted to represent both the image and text using a single image.