Does Visual Grounding Task support fine-tuning?

OFA-Sys / OFA

Official repository of OFA (ICML 2022). Paper: OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework

Apache License 2.0

2.39k stars 248 forks source link

Does Visual Grounding Task support fine-tuning? #368

Closed Practicing7 closed 1 year ago

Practicing7 commented 1 year ago

I have noticed that in your ModelScope documents, you said that OFA now only supports the training of "mnli、caption、ocr", these three tasks.

JustinLin610 commented 1 year ago

https://github.com/OFA-Sys/OFA/tree/main/run_scripts/refcoco check this directory and see if it helps

Practicing7 commented 1 year ago

Thank you! I will check this.