How to finetune the visual grounding?

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

How to finetune the visual grounding? #405

Closed kevinfu1717 closed 1 year ago

kevinfu1717 commented 1 year ago

Thanks for sharing the OFA. I want to train the visual grounding with my own data. And I would like to know how to preprocess my training data(数据格式) and how to set the configuration. Thanks!