Don't understand how compression is performed

[AlbertoPresta]

Hey,

Sorry if my question is stupid, but I do not get the code related to the actual compression of an image, concerning the atuoregressive model.
In particular, why did you put the following command at line 634 of the file google.py:

y_crop = y_crop[:, :, padding, padding]

In this way y_crop became of dimension [1,192,1,1] instead of [1,192,5,5], right?

Thanks in advance for the answer,
Alberto