I am currently training an ANN using Keras (Python3), and I am slowly optimizing the model's architecture and came across something I have not seen before.
The graph of the training and validation accuracy seems a bit odd. The graph appears 'step-like' in the sense that it is not a smooth curve, but different.
model = tf.keras.models.Sequential() model.add(Dense(7,activation='relu')) model.add(Dense(10,activation='relu')) model.add(Dense(10,activation='relu')) model.add(Dense(1,activation='sigmoid')) model.compile(optimizer=tf.keras.optimizers.Adam(0.0001), loss='binary_crossentropy', metrics=['accuracy', 'mse', 'mae']) history = model.fit( X_train.values, y_train, epochs=1000, validation_split = 0.2, verbose=1, )
My two questions are:
(1) what is this indicative of, and
(2) how do i fix this assuming it is problematic?