How to resolve "KeyError: ((1, 1, 6), '|u1')"?

[JenniferYingyiWu2020]

Hi, I have used my own data set, that is all the ".tiff" images, the "n_channels" is 3, the "n_classes" is 6. However, when I run the "python train.py --amp", the following errors appear:

Could you pls help to give me some suggestions? Thanks!

[kimile599]

what about your mask dataset, it should be single channel image

[fanxiaochen-123]

So mask is grayscale?

[fanxiaochen-123]

@JenniferYingyiWu2020 Have you solved the problem yet?