Train with Cifar data sets

[jerry3chen]

Hi, how exactly do you train with cifar10, cifar100 been private/public data sets?

In col_update code, you have to make sure the output have the same dimension, but cifar10 has 10 labels and cifar100 has 100.
If your private data is cifar10, then your model only has 10 output channels, how can you fed cifar100 in?
I am not able to find where to deal with this， would be great if you can explain/show some code.

Regards

[anywayup]

数据加载出现了好多问题，有什么大神帮忙看下

[anywayup]

FileNotFoundError: Found no valid file for the classes MNIST, cifar-100-python. Supported extensions are: .jpg, .jpeg, .png, .ppm, .bmp, .pgm, .tif, .tiff, .webp