Input size of image

[giakoumoglou]

In the training of the CNN, the input is of size (5,5,30) instead of (30,5,5) and as a result the model handles the input as an image of size 30x5 with 5 channels instead of a 5x5 image with 30 channels.