Navigation
API > API/Plugins > API/Plugins/LearningTraining
References
| Module | LearningTraining |
| Header | /Engine/Plugins/Experimental/LearningAgents/Source/LearningTraining/Public/LearningPPOTrainer.h |
| Include | #include "LearningPPOTrainer.h" |
Syntax
namespace UE
{
namespace Learning
{
enum EPPOTrainerFlags
&123;
None = 0,
UseInitialPolicyNetwork = 1 << 0,
UseInitialCriticNetwork = 1 << 1,
SynchronizeCriticNetwork = 1 << 2,
&125;
}
}
Values
| Name | Description |
|---|---|
| None | |
| UseInitialPolicyNetwork | If to send over the initial provided policy network rather than reinitialize it from random weights at the start of training. |
| UseInitialCriticNetwork | If to send over the initial provided critic network rather than reinitialize it from random weights at the start of training. |
| SynchronizeCriticNetwork | If the critic network should be synchronized between the training process and experience gathering process during training |
Remarks
PPOTrainer flags controlling some aspects of the process of communication with the trainer