Keras ImageDataGenerator shuffles the directories' labels.

[madmonl]

I am working on a project which involves the use of a datagen. The project's data consists of data arranged in subclasses that resides in directories named: '1', '2', ..., '64'.
When i checked the results on the test set, I recognized that the datagen rearanged the directories in a non-lexicographic way using a directory as follows:
dict = {'10':0, '11':1, .... , '7':63}. (under the field: validation_generator.indices).
Meaning that if my probabilities vector of size 64, looks like the following:
[1, 0, 0, ... , 0]
The network will consider the samples as class '10'-th, while i want it rearranged in a lexicographic way, meaning the above probabilities vector would be the output of an image belonging to the '0'-th class.
Is there a way of controlling the datagen's labels shuffle system?

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed after 30 days if no further activity occurs, but feel free to re-open a closed issue if needed.