Maybe I'm getting it wrong, but why is the model always trained on the full training set.
When I use verbose=2, every epoch is using the whole training set provided by the data function and it looks like the validation split in the fit function is not used.
I would appreciate any explenation!
Maybe I'm getting it wrong, but why is the model always trained on the full training set. When I use verbose=2, every epoch is using the whole training set provided by the data function and it looks like the validation split in the fit function is not used. I would appreciate any explenation!