No module named 'models'

[zsp1993]

HI,
when I run PENCIL.py, there was an import error said no module named 'models',
look forward to see your complete code.

[zsp1993]

another issue is in self.train_data = self.train_data.reshape((45000, 3, 32, 32)),
which result in ValueError: cannot reshape array of size 153600000 into shape (45000,3,32,32)

[mily33]

(1)You can modify models in torchvision.model_zoo to build your own models.

(2)153600000=50000332*32. You should retain 10% of the CIFAR-10 training data (i.e. 5000 images) as the validation set. I split the data_batch_5 into train and val dataset, and it works for me.

@zsp1993 Hope it helps!