About Dataset reason_seg/explanatory

[asad14053]

In train_ds.py, there is code for train validation on reason_seg data. But, where and how was the train/ validation/ testing on reason_seg/explanatory?
a) Could you please share some insights on how LISA-7B-explanatory is trained differently from regular LISA-7B? What was extra added to the explanatory model?

parser.add_argument("--vqa_data", default="llava_instruct_150k", type=str)
parser.add_argument("--reason_seg_data", default="ReasonSeg|train", type=str)
parser.add_argument("--val_dataset", default="ReasonSeg|val", type=str)