Request on image-to-text implementation

[herry123435]

Dear Dr. He,

Thank you for releasing the text-to-image code and checkpoints.
This repository has been very helpful for understanding and reproducing your work.

I had a question about the image-to-text direction. Would it be possible to also release the image-to-text implementation (training/inference code, configs, and checkpoints), if available?

I would greatly appreciate any guidance on:

• whether image-to-text code/checkpoints are planned for release,
• expected timeline (if possible),
• and, if not, what minimum components should be added to the current repo to reproduce image-to-text results.

Thanks again for your great work and for open-sourcing the project.

Best regards,