How can i train a model on commonly used MIML dataset like scene and text

[heivens]

I want to train a model on the multi-instance multi-label dataset (scene, text), but it does not optimize. So i wonder if there is a demo that i can borrow to bulid a model on scene and text datasets