actor.py

from keras import layers, models
from keras import backend as K
from keras import optimizers


class Actor:
    """Actor (Policy) Model."""

    def __init__(self, state_size, action_size, action_high, action_low):
        """Initialize Actor instance."""
        self.state_size = state_size
        self.action_size = action_size
        self.action_high = action_high
        self.action_low = action_low
        self.action_range = self.action_high - self.action_low

        self.build_model()

    def build_model(self):
        # input layer (states)
        states = layers.Input(shape=(self.state_size,), name='states')

        # hidden layers
        net = layers.Dense(units=32, activation='relu')(states)
        net = layers.Dense(units=64, activation='relu')(net)
        net = layers.Dense(units=32, activation='relu')(net)

        # final layer with sigmoid
        raw_actions = layers.Dense(units=self.action_size, activation='sigmoid',
                                   name='raw_actions')(net)

        # Scale output for each action dimension to proper range
        actions = layers.Lambda(lambda x: (x * self.action_range) + self.action_low,
                                name='actions')(raw_actions)

        # Keras model
        self.model = models.Model(inputs=states, outputs=actions)

        # loss function
        action_gradients = layers.Input(shape=(self.action_size,))
        loss = K.mean(-action_gradients * actions)

        # optimizer and training function
        optimizer = optimizers.Adam()
        updates_op = optimizer.get_updates(
            params=self.model.trainable_weights, loss=loss)
        self.train_fn = K.function(
            inputs=[self.model.input, action_gradients, K.learning_phase()],
            outputs=[],
            updates=updates_op)