Zero Accuracy DNN

Question

I have a deep linear regression model that I have been working on

self.model = keras.Sequential([
            keras.layers.Dense(64, activation=tf.nn.selu),
            keras.layers.Dropout(0.4),
            keras.layers.Dense(64, activation=tf.nn.selu),
            keras.layers.Dropout(0.4),
            keras.layers.Dense(1)

I compiled it with the mean absolute error

self.model.compile(
            optimizer=tf.optimizers.Adam(learning_rate=self.learning_rate),
            loss=tf.losses.MeanAbsoluteError(),metrics=['accuracy'])

And then I trained it

self.model.fit(self.X_train, self.y_train, epochs=epochs, batch_size=batch_size)

However I keep getting accuracy as 0. Is there something wrong I am doing?

Zero Accuracy DNN

0 Answers