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
Can visual-grounding be changed to one-to-many target positioning? #400
@DingLing113 Currently OFA doesn't support this feature.