Any missing classes/functions in viclip_text.py and viclip_vision.py?

[lixin4ever]

Hello, I am so glad that you open-sourced the checkpoints and the demo script recently. When I run the provided demo script, I found that the viclip.py attempt to import clip_joint_l14, clip_joint_b16 and clip_text_l14, clip_text_b16, which are not pre-defined in viclip_text.py or viclip_vision.py. Just wonder if the current files are the ones in your local run?

(Also, the demo script can not run normally due to circular import, but it doesn't matter, can be fixed easily

[shepnerd]

viclip_text.py has been updated, while the viclip_vision.py file appears to be functioning properly. We encourage you to test them and report any further issues you encounter.

[lixin4ever]

Great! Thanks for your prompt update.

[lixin4ever]

Already tested, it perfectly works.

viclip_text.py has been updated, while the viclip_vision.py file appears to be functioning properly. We encourage you to test them and report any further issues you encounter.

Just one minor question, why is the best-performed ViCLIP-L checkpoint (i.e., ViCLIP-L + InternVid-10M-FLT) much smaller than other ViCLIP-Ls in terms of file size?