Using data.Iterator.get_next() with horovod

Question

I try to parallelize some institute\'s tensorflow code with horovod to work in a distributed setup.

I create the dataset and split it for each horovod process using