class probability in segmentation

[DoDucNhan]

Hello everyone. I use the BEiT network for semantic segmentation and have trouble determining if my model segments the object correctly. In this case, there won't be any ground truth to validate. The result return from "inference_segmentor" only one array with the class id.

Is there any way I can obtain class probability or confidence level which tells how confident a model is about that segmentation like we get in rcnn?