Adding expected input data section
Currently in the training frame we give "data examples", but in the case of CIFAR, it's just showing how images look like visually. I added some temporary information of how the data should be formatted (As raised by Martin in a previous issue)
This is however important so we should make this section look cleaner.
@nacho @martinjaggi I suggest we add this important information to either the Description or Dataset Input section.
What would be nice is a preview of an example image against the preview of one of the user's given images, with the required dimensions clearly stated. For CSV datasets, to simply display the expected columns is enough.
done in #389