Theoretically and practically, are the hidden layers of MLPclassifier (refer to hidden_layer_sizes)
mlp = MLPClassifier(hidden_layer_sizes=(4, 3, 2, 1), max_iter = 100, activation = 'relu', solver = 'adam', verbose = type_spec_from_value, random_state = 100, learning_rate = 'invscaling', early_stopping=False )
the same as the Dense layers of tensorflow/keras
mlp = Sequential() mlp.add(Dense(4)) mlp.add(Dense(3)) mlp.add(Dense(2)) mlp.add(Dense(3))
Yes, they are the same. In both cases, the parameters specify the number of neurons.
Answered By – lua