Thanks for your work, I'm trying to use FasterSeg with my own custom data. My tack is just Two-class task.
The imgsize of my data is 1280*640, and the integers in mask are 0 and 15, 0 represents background and 15 is the target. The "dataset files" are also done:
The num_classes I changed to 2, height and width are also adjusted.
Error:
Thanks for your work, I'm trying to use FasterSeg with my own custom data. My tack is just Two-class task. The imgsize of my data is 1280*640, and the integers in mask are 0 and 15, 0 represents background and 15 is the target. The "dataset files" are also done:
The num_classes I changed to 2, height and width are also adjusted.
Error:
![image](https://user-images.githubusercontent.com/58512513/122226940-04503800-cee9-11eb-8069-f486a7240713.png)