Hello,
I'm trying to train this model by using my own dataset.
But how could I build my own coco-like panoptic segmentation dataset with my own dataset to use this code?
I wrote a script for the sunrgb-d and nyuv2 dataset to prepare the annotation files and ground truth pngs from the instance/semantic mask pngs. Message me if still interessed.
Hello, I'm trying to train this model by using my own dataset. But how could I build my own coco-like panoptic segmentation dataset with my own dataset to use this code?