Module: tf_agents.networks

Networks Module.


actor_distribution_network module: Sample Keras actor network that generates distributions.

actor_distribution_rnn_network module: Sample Keras actor network with LSTM cells that generates distributions.

categorical_projection_network module: Project inputs to a categorical distribution object.

categorical_q_network module: A Q-network for categorical DQN.

encoding_network module: Keras Encoding Network.

expand_dims_layer module: Keras layer performing the equivalent of tf.expand_dims.

layer_utils module: Network layer utilities.

lstm_encoding_network module: Keras LSTM Encoding Network.

mask_splitter_network module: Wrapper network that handles action constraint portion of the observation.

nest_map module: Network layer that allows mapping multiple inputs.

network module: Base extension to Keras network to simplify copy operations.

normal_projection_network module: Project inputs to a normal distribution object.

q_network module: Sample Keras networks for DQN.

q_rnn_network module: Sample recurrent Keras network for DQN.

sequential module: Keras layer to replace the Sequential Model object.

utils module: Network utilities.

value_network module: Sample Keras Value Network.

value_rnn_network module: Sample Keras Value Network with LSTM cells .


class NestMap: The NestMap network processes nested inputs via nested layers.

class Network: A class used to represent networks used by TF-Agents policies and agents.

class Sequential: The Sequential Network represents a sequence of Keras layers.


NestFlatten(...): Returns a Keras layer that takes a nest of inputs, and returns a list.