garage.tf.q_functions package

Q-Functions for TensorFlow-based algorithms.

class QFunction(name)[source]

Bases: abc.ABC

Q-function base class without Parameterzied.

Parameters:name (str) – Name of the Q-fucntion, also the variable scope.
clone(name)[source]

Return a clone of the Q-function.

It should only copy the configuration of the Q-function, not the parameters.

Parameters:name (str) – Name of the newly created q-function.
get_global_vars()[source]

Get all global variables under the QFunction scope.

get_qval_sym(*input_phs)[source]

Symbolic graph for q-network.

All derived classes should implement this function.

Parameters:input_phs (list[tf.Tensor]) – Recommended to be positional arguments, e.g. def get_qval_sym(self, state_input, action_input).
get_regularizable_vars()[source]

Get all network weight variables under the QFunction scope.

get_trainable_vars()[source]

Get all trainable variables under the QFunction scope.

log_diagnostics(paths)[source]

Log extra information per iteration based on the collected paths.

class ContinuousMLPQFunction(env_spec, name='ContinuousMLPQFunction', hidden_sizes=(32, 32), action_merge_layer=-2, hidden_nonlinearity=<function relu>, hidden_w_init=<tensorflow.python.ops.init_ops.GlorotUniform object>, hidden_b_init=<tensorflow.python.ops.init_ops.Zeros object>, output_nonlinearity=None, output_w_init=<tensorflow.python.ops.init_ops.GlorotUniform object>, output_b_init=<tensorflow.python.ops.init_ops.Zeros object>, input_include_goal=False, layer_normalization=False)[source]

Bases: garage.tf.q_functions.base.QFunction

Continuous MLP QFunction.

This class implements a q value network to predict q based on the input state and action. It uses an MLP to fit the function of Q(s, a).

Parameters:
  • env_spec (garage.envs.env_spec.EnvSpec) – Environment specification.
  • name (str) – Name of the q-function, also serves as the variable scope.
  • hidden_sizes (list[int]) – Output dimension of dense layer(s). For example, (32, 32) means the MLP of this q-function consists of two hidden layers, each with 32 hidden units.
  • action_merge_layer (int) – The index of layers at which to concatenate action inputs with the network. The indexing works like standard python list indexing. Index of 0 refers to the input layer (observation input) while an index of -1 points to the last hidden layer. Default parameter points to second layer from the end.
  • hidden_nonlinearity (callable) – Activation function for intermediate dense layer(s). It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • hidden_w_init (callable) – Initializer function for the weight of intermediate dense layer(s). The function should return a tf.Tensor.
  • hidden_b_init (callable) – Initializer function for the bias of intermediate dense layer(s). The function should return a tf.Tensor.
  • output_nonlinearity (callable) – Activation function for output dense layer. It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • output_w_init (callable) – Initializer function for the weight of output dense layer(s). The function should return a tf.Tensor.
  • output_b_init (callable) – Initializer function for the bias of output dense layer(s). The function should return a tf.Tensor.
  • input_include_goal (bool) – Whether input includes goal.
  • layer_normalization (bool) – Bool for using layer normalization.
clone(name)[source]

Return a clone of the Q-function.

It only copies the configuration of the Q-function, not the parameters.

Parameters:name (str) – Name of the newly created q-function.
Returns:A new instance with same arguments.
Return type:ContinuousMLPQFunction
get_qval(observation, action)[source]

Q Value of the network.

Parameters:
  • observation (np.ndarray) – Observation input.
  • action (np.ndarray) – Action input.
Returns:

Q values.
Return type:

np.ndarray
get_qval_sym(state_input, action_input, name)[source]

Symbolic graph for q-network.

Parameters:
  • state_input (tf.Tensor) – The state input tf.Tensor to the network.
  • action_input (tf.Tensor) – The action input tf.Tensor to the network.
  • name (str) – Network variable scope.
Returns:

The output of Discrete MLP QFunction.
Return type:

tf.Tensor
inputs

Return the input tensor.

Returns:The input tensors of the model.
Return type:tf.Tensor
class DiscreteCNNQFunction(env_spec, filter_dims, num_filters, strides, hidden_sizes=[256], name=None, padding='SAME', max_pooling=False, pool_strides=(2, 2), pool_shapes=(2, 2), cnn_hidden_nonlinearity=<function relu>, hidden_nonlinearity=<function relu>, hidden_w_init=<tensorflow.python.ops.init_ops.GlorotUniform object>, hidden_b_init=<tensorflow.python.ops.init_ops.Zeros object>, output_nonlinearity=None, output_w_init=<tensorflow.python.ops.init_ops.GlorotUniform object>, output_b_init=<tensorflow.python.ops.init_ops.Zeros object>, dueling=False, layer_normalization=False)[source]

Bases: garage.tf.q_functions.base.QFunction

Q function based on a CNN-MLP structure for discrete action space.

This class implements a Q value network to predict Q based on the input state and action. It uses an CNN and a MLP to fit the function of Q(s, a).

Parameters:
  • env_spec (garage.envs.env_spec.EnvSpec) – Environment specification.
  • filter_dims (tuple[int]) – Dimension of the filters. For example, (3, 5) means there are two convolutional layers. The filter for first layer is of dimension (3 x 3) and the second one is of dimension (5 x 5).
  • num_filters (tuple[int]) – Number of filters. For example, (3, 32) means there are two convolutional layers. The filter for the first layer has 3 channels and the second one with 32 channels.
  • strides (tuple[int]) – The stride of the sliding window. For example, (1, 2) means there are two convolutional layers. The stride of the filter for first layer is 1 and that of the second layer is 2.
  • hidden_sizes (list[int]) – Output dimension of dense layer(s). For example, (32, 32) means the MLP of this q-function consists of two hidden layers, each with 32 hidden units.
  • name (str) – Variable scope of the cnn.
  • padding (str) – The type of padding algorithm to use, either ‘SAME’ or ‘VALID’.
  • max_pooling (bool) – Boolean for using max pooling layer or not.
  • pool_shapes (tuple[int]) – Dimension of the pooling layer(s). For example, (2, 2) means that all the pooling layers have shape (2, 2).
  • pool_strides (tuple[int]) – The strides of the pooling layer(s). For example, (2, 2) means that all the pooling layers have strides (2, 2).
  • cnn_hidden_nonlinearity (callable) – Activation function for intermediate dense layer(s) in the CNN. It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • hidden_nonlinearity (callable) – Activation function for intermediate dense layer(s) in the MLP. It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • hidden_w_init (callable) – Initializer function for the weight of intermediate dense layer(s) in the MLP. The function should return a tf.Tensor.
  • hidden_b_init (callable) – Initializer function for the bias of intermediate dense layer(s) in the MLP. The function should return a tf.Tensor.
  • output_nonlinearity (callable) – Activation function for output dense layer in the MLP. It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • output_w_init (callable) – Initializer function for the weight of output dense layer(s) in the MLP. The function should return a tf.Tensor.
  • output_b_init (callable) – Initializer function for the bias of output dense layer(s) in the MLP. The function should return a tf.Tensor.
  • dueling (bool) – Bool for using dueling network or not.
  • layer_normalization (bool) – Bool for using layer normalization or not.
clone(name)[source]

Return a clone of the Q-function.

It only copies the configuration of the Q-function, not the parameters.

Parameters:name – Name of the newly created q-function.
get_qval_sym(state_input, name)[source]

Symbolic graph for q-network.

Parameters:
  • state_input (tf.Tensor) – The state input tf.Tensor to the network.
  • name (str) – Network variable scope.
Returns:

The tf.Tensor output of Discrete CNN QFunction.
input

Input tf.Tensor of the Q-function.

q_vals

Q values.

class DiscreteMLPQFunction(env_spec, name=None, hidden_sizes=(32, 32), hidden_nonlinearity=<function relu>, hidden_w_init=<tensorflow.python.ops.init_ops.GlorotUniform object>, hidden_b_init=<tensorflow.python.ops.init_ops.Zeros object>, output_nonlinearity=None, output_w_init=<tensorflow.python.ops.init_ops.GlorotUniform object>, output_b_init=<tensorflow.python.ops.init_ops.Zeros object>, dueling=False, layer_normalization=False)[source]

Bases: garage.tf.q_functions.base.QFunction

Discrete MLP Q Function.

This class implements a Q-value network. It predicts Q-value based on the input state and action. It uses an MLP to fit the function Q(s, a).

Parameters:
  • env_spec (garage.envs.env_spec.EnvSpec) – Environment specification.
  • name (str) – Name of the q-function, also serves as the variable scope.
  • hidden_sizes (list[int]) – Output dimension of dense layer(s). For example, (32, 32) means the MLP of this q-function consists of two hidden layers, each with 32 hidden units.
  • hidden_nonlinearity (callable) – Activation function for intermediate dense layer(s). It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • hidden_w_init (callable) – Initializer function for the weight of intermediate dense layer(s). The function should return a tf.Tensor.
  • hidden_b_init (callable) – Initializer function for the bias of intermediate dense layer(s). The function should return a tf.Tensor.
  • output_nonlinearity (callable) – Activation function for output dense layer. It should return a tf.Tensor. Set it to None to maintain a linear activation.
  • output_w_init (callable) – Initializer function for the weight of output dense layer(s). The function should return a tf.Tensor.
  • output_b_init (callable) – Initializer function for the bias of output dense layer(s). The function should return a tf.Tensor.
  • layer_normalization (bool) – Bool for using layer normalization.
clone(name)[source]

Return a clone of the Q-function.

It only copies the configuration of the Q-function, not the parameters.

Parameters:name (str) – Name of the newly created q-function.
get_qval_sym(state_input, name)[source]

Symbolic graph for q-network.

Parameters:
  • state_input (tf.Tensor) – The state input tf.Tensor to the network.
  • name (str) – Network variable scope.
Returns:

The tf.Tensor output of Discrete MLP QFunction.
input

Get input.

q_vals

Return the Q values, the output of the network.