Functions — github.com/artem-oppermann/Deep-Reinforcement-Learning

hub / github.com/artem-oppermann/Deep-Reinforcement-Learning / functions

Functions72 in github.com/artem-oppermann/Deep-Reinforcement-Learning

⨍ Functions 72 ◇ Types & classes 15

↓ 21 callersMethodrun

(self)

src/q learning/cartpole_env.py:64

↓ 7 callersMethodrun

Main loop for the running of the episodes.

src/double q learning/cartpole_env.py:87

↓ 4 callersMethodfeaturize_state

Returns the featurized representation for a state.

src/policy gradients/stochastic/stochastic_pg.py:202

↓ 4 callersMethodset_session

Set the session

src/policy gradients/deterministic/ddpg.py:91

↓ 3 callersMethodget_action

Get an action for a certain state

src/policy gradients/deterministic/ddpg.py:103

↓ 3 callersMethodget_mountain_env

Getter function for the OpenAI Gym instance

src/policy gradients/stochastic/stochastic_pg.py:194

↓ 2 callersMethodaction_estimator

Build the neural network that estimates the action for a given state

src/policy gradients/deterministic/ddpg.py:68

↓ 2 callersMethodaction_value_estimator

Build the neural network that estimates the action-values

src/policy gradients/deterministic/ddpg.py:174

↓ 2 callersMethodget_state_size

Getter function for the state-size in the environment

src/policy gradients/stochastic/stochastic_pg.py:198

↓ 2 callersMethodinit_target_network

Initialize the parameters of the target-network

src/policy gradients/deterministic/ddpg.py:95

↓ 2 callersMethodpredict

Predict the state-value for the state :param state: Current state in the OpenAI Gym Environment

src/policy gradients/stochastic/stochastic_pg.py:143

↓ 2 callersMethodset_session

Sets the session of the appropriate network.

src/double q learning/ddqn_model.py:175

↓ 2 callersMethodset_session

Setter method for the session

src/policy gradients/stochastic/stochastic_pg.py:98

↓ 2 callersMethodset_session

(self, session)

src/q learning/dqn_model.py:134

↓ 2 callersMethodtrain

Train the actor network

src/policy gradients/deterministic/ddpg.py:107

↓ 2 callersMethodupdate

Update operation for the weights

src/policy gradients/stochastic/stochastic_pg.py:83

↓ 2 callersMethodupdate_target_parameter

Update the parameters of the target-network

src/policy gradients/deterministic/ddpg.py:99

↓ 1 callersMethod_mu_classifier

Linear clsasifier for the mean of the gaussian distribution

src/policy gradients/stochastic/stochastic_pg.py:64

↓ 1 callersMethod_sigma_classifier

Linear clsasifier for the variance of the gaussian distribution

src/policy gradients/stochastic/stochastic_pg.py:73

↓ 1 callersMethod_value_estimator

Linear classifier for the estimatation of the state-value

src/policy gradients/stochastic/stochastic_pg.py:130

↓ 1 callersMethodaddExperience

Add experience to the memory. :param state: current state s, given by the environment :param action: action taken in state s

src/double q learning/exp_replay.py:29

↓ 1 callersMethodaddExperience

(self, state, action, reward, next_state,done)

src/policy gradients/deterministic/experience_replay.py:20

↓ 1 callersMethodaddExperience

(self, state, action, reward, next_state,done)

src/q learning/exp_replay.py:26

↓ 1 callersMethodbuild_q_network

Build the Q-Network

src/double q learning/ddqn_model.py:98

↓ 1 callersMethodbuild_q_network

(self)

src/q learning/dqn_model.py:74

↓ 1 callersMethodbuild_target_network

Build the Target-Network

src/double q learning/ddqn_model.py:78

↓ 1 callersMethodbuild_target_network

(self)

src/q learning/dqn_model.py:55

↓ 1 callersMethodcalculate_Q

Compute the action-value

src/policy gradients/deterministic/ddpg.py:210

↓ 1 callersMethodcompute_gradients

Compute the gradients of the action_value estimator neural network

src/policy gradients/deterministic/ddpg.py:197

↓ 1 callersMethodget_action

Calcualte Q(s,a) by the Q-Network, use the epislon-greedy policy to pick an action.. :param s: current state s, given by the environment

src/double q learning/ddqn_model.py:163

↓ 1 callersMethodget_action

(self, X, eps)

src/q learning/dqn_model.py:126

↓ 1 callersMethodget_experience

(self)

src/double q learning/exp_replay.py:20

↓ 1 callersMethodget_experience

(self)

src/q learning/exp_replay.py:16

↓ 1 callersMethodget_min_experience_count

(self)

src/double q learning/exp_replay.py:23

↓ 1 callersMethodget_min_experience_count

(self)

src/q learning/exp_replay.py:19

↓ 1 callersMethodget_noise

(self, ep)

src/policy gradients/deterministic/noise.py:25

↓ 1 callersMethodplayEpisode

Play one single episode of the environmet. :param eps: value of probability epsilon

src/double q learning/cartpole_env.py:46

↓ 1 callersMethodplayEpisode

Play an episode in the environment

src/policy gradients/deterministic/ddpg.py:317

↓ 1 callersMethodplayEpisode

Play an episode in the OpenAI Gym :param episode: Number of the current episode

src/policy gradients/stochastic/stochastic_pg.py:241

↓ 1 callersMethodplayEpisode

(self,eps)

src/q learning/cartpole_env.py:31

↓ 1 callersMethodrun_model

Main loop. Runs the environment and traing the networks

src/policy gradients/deterministic/ddpg.py:347

↓ 1 callersMethodrun_model

Run the environment for a particular number of episodes.

src/policy gradients/stochastic/stochastic_pg.py:299

↓ 1 callersMethodsample_action

Sample an action from the gaussian distribution

src/policy gradients/stochastic/stochastic_pg.py:91

↓ 1 callersMethodtrain_networks

Training of the actor and critic networks

src/policy gradients/deterministic/ddpg.py:287

↓ 1 callersMethodtrain_q_network

Train the Q-Network.

src/double q learning/ddqn_model.py:117

↓ 1 callersMethodtrain_q_network

(self)

src/q learning/dqn_model.py:92

↓ 1 callersMethodupdate_target_parameter

Set the parameter of the target network to the parameter of the Q-network

src/double q learning/ddqn_model.py:180

↓ 1 callersMethodupdate_target_parameter

(self)

src/q learning/dqn_model.py:138

Method__init__

This class build a model for either a target or Q-Network and impliments the methods of Deep Double Q-Learning. :pa

src/double q learning/ddqn_model.py:13

Method__init__

(self)

src/double q learning/exp_replay.py:14

Method__init__

(self, FLAGS)

src/double q learning/cartpole_env.py:23

Method__init__

This class implements the actor for the deterministic policy gradients model. The actor class determines the action that the agent mu

src/policy gradients/deterministic/ddpg.py:18

Method__init__

This class implements the Critic for the stochastic policy gradient model. The critic provides a state-value for the current state en

src/policy gradients/deterministic/ddpg.py:118

Method__init__

This class build the model that implements the deterministic gradient descent algorithm. :param FLAGS: TensorFlow f

src/policy gradients/deterministic/ddpg.py:248

Method__init__

(self, max_experience,min_experience, flags)

src/policy gradients/deterministic/experience_replay.py:11

Method__init__

(self, env,mu= 0.0, sigma=0.2, theta=.15, dt=1e-2, x0=None)

src/policy gradients/deterministic/noise.py:12

Method__init__

This class implements the actor for the stochastic policy gradient model. The actor class determines the action that the agent must t

src/policy gradients/stochastic/stochastic_pg.py:20

Method__init__

This class implements the Critic for the stochastic policy gradient model. The critic provides a state-value for the current state en

src/policy gradients/stochastic/stochastic_pg.py:105

Method__init__

This class implements the OpenAI Gym Environment for MountainCarContinuous-v0

src/policy gradients/stochastic/stochastic_pg.py:170

Method__init__

This class build the model that implements the stochstic gradient descent algorithm. :param FLAGS: TensorFlow flags

src/policy gradients/stochastic/stochastic_pg.py:213

Method__init__

(self)

src/q learning/exp_replay.py:10

Method__init__

(self, FLAGS)

src/q learning/cartpole_env.py:10

Method__init__

(self, scope, env, target_network, flags, exp_replay)

src/q learning/dqn_model.py:8

Methodget_max_experience_count

(self)

src/double q learning/exp_replay.py:26

Methodget_max_experience_count

(self)

src/q learning/exp_replay.py:22

Methodget_sample

(self)

src/policy gradients/deterministic/experience_replay.py:36

Methodinit_target_network

Initialize the parameters of the target-network

src/policy gradients/deterministic/ddpg.py:236

Methodset_session

Set the session

src/policy gradients/deterministic/ddpg.py:231

Methodset_session

Setter method for the session

src/policy gradients/stochastic/stochastic_pg.py:139

Methodtrain

Train the actor network

src/policy gradients/deterministic/ddpg.py:221

Methodupdate

Update the weights of the critic :param state: Current state in the OpenAI Gym Environment :param target: TD-Target

src/policy gradients/stochastic/stochastic_pg.py:153

Methodupdate_target_parameter

Update the parameters of the target-network

src/policy gradients/deterministic/ddpg.py:241