https://mooseframework.inl.gov
Loading...
Searching...
No Matches
Public Attributes | List of all members
LibtorchRLTrajectoryBuffer::RewardStatistics Struct Reference

Reward statistics over the stored trajectories. More...

#include <LibtorchRLTrajectoryBuffer.h>

Public Attributes

Real mean = 0.0
 Mean reward over every stored transition.
 
Real std = 0.0
 Standard deviation of every stored transition reward about the overall mean.
 
std::vector< Real > trajectory_means
 Mean reward of each non-empty trajectory.
 
std::vector< Real > trajectory_stds
 Reward standard deviation of each non-empty trajectory about its own mean.
 

Detailed Description

Reward statistics over the stored trajectories.

Definition at line 79 of file LibtorchRLTrajectoryBuffer.h.

Member Data Documentation

◆ mean

Real LibtorchRLTrajectoryBuffer::RewardStatistics::mean = 0.0

Mean reward over every stored transition.

Definition at line 82 of file LibtorchRLTrajectoryBuffer.h.

Referenced by LibtorchDRLControlTrainer::computeEpisodeRewardStatistics(), and LibtorchRLTrajectoryBuffer::rewardStatistics().

◆ std

Real LibtorchRLTrajectoryBuffer::RewardStatistics::std = 0.0

Standard deviation of every stored transition reward about the overall mean.

Definition at line 84 of file LibtorchRLTrajectoryBuffer.h.

Referenced by LibtorchRLTrajectoryBuffer::rewardStatistics().

◆ trajectory_means

std::vector<Real> LibtorchRLTrajectoryBuffer::RewardStatistics::trajectory_means

Mean reward of each non-empty trajectory.

Definition at line 86 of file LibtorchRLTrajectoryBuffer.h.

Referenced by LibtorchRLTrajectoryBuffer::rewardStatistics().

◆ trajectory_stds

std::vector<Real> LibtorchRLTrajectoryBuffer::RewardStatistics::trajectory_stds

Reward standard deviation of each non-empty trajectory about its own mean.

Definition at line 88 of file LibtorchRLTrajectoryBuffer.h.

Referenced by LibtorchRLTrajectoryBuffer::rewardStatistics().


The documentation for this struct was generated from the following file: