语义分割训练过程中的num_classes是8还是7 ？

[Voruarn]

因为我用 cross_entropy loss训练语义分割模型时，num_classes为7会报错，num_classes为8可以正常运行。 但是Dataset and Contest部分却说 And the no-data regions were assigned 0 which should be ignored. 有点困惑。

[Junjue-Wang]

num_classes=7,我的dataloader自动把0变成了负1 ignore了

[Voruarn]

好的，谢谢