OpenGVLab/LAMM

:question: Could you specify all multitasks Octavius was trained on?

simon-lund opened this issue · 0 comments

In the Octavius paper only a few tasks are listed and the listing is abbreviated by "...and other common tasks".

Could you please specify all tasks and also confirm if you ONLY used COCO and Bamboo images from LammV2 as mentioned in the paper.

Thank you very much 👍