Learner with multiple datasets

Hi, I have a model that outputs a mask and class probabilities. I used SegmentationItemList for only mask and MultiCategoryList for class probabilities vector, the problem is how to give these two datasets to Learner so that the loss_func receives two targets one from each dataset. Outputs for the loss_func comes from model. and compute. Also is there a way to combine SegmentationItemList and MultiCategoryList into single data such that data.show_batch, learn.show_results will plot both mask along with categories on top?