This repository was archived by the owner on Oct 14, 2025. It is now read-only.

Description
We had deactivated shuffling of the dataset prior to tf.Dataset introduction because of tf's shuffle-repeat "bug".
Since tf.Dataset allows to first shuffle and then repeat, we may reactivate shuffling by default. This needs checking whether test fail because of different summation orders and so on.
Note that this should in principle change nothing of batch_size equals the dimension of the dataset.