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

OpenGVLab / LAMM

[NeurIPS 2023 Datasets and Benchmarks Track] LAMM: Multi-Modal Large Language Models and Applications as AI Agents

https://openlamm.github.io/

286 stars 15 forks source link

Open simon-lund opened 7 months ago

simon-lund commented 7 months ago

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 :+1: