Hi everyone,
I am trying to train a model under NengoDL simulator, my dataset is generating batches using Keras sequence generator (keras.utils.data_utils.Sequence).
According to the documentation, the fit function (sim.fit) can take argument (x) as a generator. The issue is that Keras will not be able to train a model that uses a generator with a pre-determined batch_size (should be None), and the initializer of the simulator sets the batch_size to 1 if we pass None to the minibatch_size argument while creating the simulator (line 446 in simulator.py link).
Given the information above, I am getting a ValueError because of the minibatch_size not being None. I get that error from (training_v1.py line 1893 link).
Am I missing something that I should be doing when using generators? or is there a work around this issue?
Many thanks in advance!
Nidal