When adv_level was set to "input", get_cls_parameter() could not get self.hidden_layers.parameters() as the hidden layers were not initialised yet, throwing
AttributeError: 'MLP' object has no attribute 'hidden_layers'.
I've changed the code so that the hidden layer init happens earlier.
When adv_level was set to "input",
get_cls_parameter()could not getself.hidden_layers.parameters()as the hidden layers were not initialised yet, throwingAttributeError: 'MLP' object has no attribute 'hidden_layers'.I've changed the code so that the hidden layer init happens earlier.