TD3Options<T>
Options & Configuration · AiDotNet.Models.Options
Configuration options for TD3 agent.
Inheritance: ModelOptions → ReinforcementLearningOptions<T> → TD3Options<T>
Constructors
| Constructor | Summary |
|---|---|
TD3Options() |
Creates options with the default values. |
Properties
| Property | Summary |
|---|---|
ILossFunction<T> CriticLossFunction { get; set; } |
Loss function used to train the critic network. |
List<int> ActorHiddenLayers { get; set; } |
Hidden layer sizes of the actor network. |
List<int> CriticHiddenLayers { get; set; } |
Hidden layer sizes of the critic network. |
T ActorLearningRate { get; set; } |
Learning rate of the actor network. |
T CriticLearningRate { get; set; } |
Learning rate of the critic network. |
T TargetUpdateTau { get; set; } |
Soft-update coefficient for the target networks: each update moves them this fraction of the way toward the online networks. |
double ExplorationNoise { get; set; } |
Standard deviation of the Gaussian noise added to actions for exploration. |
double TargetNoiseClip { get; set; } |
Clip range of the noise added to target actions (target policy smoothing). |
double TargetPolicyNoise { get; set; } |
Standard deviation of the noise added to target actions (target policy smoothing). |
int ActionSize { get; set; } |
Number of discrete actions, or the dimension of the action vector for a continuous action space. |
int PolicyUpdateFrequency { get; set; } |
Number of critic updates per actor and target update (delayed policy updates). |
int StateSize { get; set; } |
Number of features in one observation of the environment state. |