Navigation
API > API/Plugins > API/Plugins/LearningTraining
Large buffer that sequentially concatenates a series of episodes in a large flat array. Used to collate episodic data together from multiple instances.
| Name | FReplayBuffer |
| Type | struct |
| Header File | /Engine/Plugins/Experimental/LearningAgents/Source/LearningTraining/Public/LearningExperience.h |
| Include Path | #include "LearningExperience.h" |
Syntax
struct FReplayBuffer
Variables
Protected
| Name | Type | Remarks | Include Path | Unreal Specifiers |
|---|---|---|---|---|
| ActionModifierNames | TArray< FName, TInlineAllocator< 1 > > | LearningExperience.h | ||
| ActionModifiers | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| ActionModifierSchemaIds | TArray< int32, TInlineAllocator< 1 > > | LearningExperience.h | ||
| ActionNames | TArray< FName, TInlineAllocator< 1 > > | LearningExperience.h | ||
| Actions | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| ActionSchemaIds | TArray< int32, TInlineAllocator< 1 > > | LearningExperience.h | ||
| bHasCompletions | bool | LearningExperience.h | ||
| bHasFinalMemoryStates | bool | LearningExperience.h | ||
| bHasFinalObservations | bool | LearningExperience.h | ||
| EpisodeCompletionModes | TLearningArray< 1, ECompletionMode > | LearningExperience.h | ||
| EpisodeFinalMemoryStates | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| EpisodeFinalObservations | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| EpisodeLengths | TLearningArray< 1, int32 > | LearningExperience.h | ||
| EpisodeNum | int32 | LearningExperience.h | ||
| EpisodeStarts | TLearningArray< 1, int32 > | LearningExperience.h | ||
| MaxEpisodeNum | int32 | LearningExperience.h | ||
| MaxStepNum | int32 | LearningExperience.h | ||
| MemoryStateNames | TArray< FName, TInlineAllocator< 1 > > | LearningExperience.h | ||
| MemoryStates | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| ObservationNames | TArray< FName, TInlineAllocator< 1 > > | LearningExperience.h | ||
| Observations | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| ObservationSchemaIds | TArray< int32, TInlineAllocator< 1 > > | LearningExperience.h | ||
| RewardNames | TArray< FName, TInlineAllocator< 1 > > | LearningExperience.h | ||
| Rewards | TArray< TLearningArray< 2, float >, TInlineAllocator< 1 > > | LearningExperience.h | ||
| StepNum | int32 | LearningExperience.h |
Functions
Public
| Name | Remarks | Include Path | Unreal Specifiers |
|---|---|---|---|
bool AddEpisodes
(
const TLearningArrayView< 1, const ECompletionMode > InEpisodeCompletionModes, |
Add a set of episodes to the replay buffer | LearningExperience.h | |
void AddRecords
(
const int32 InEpisodeNum, |
Alternate way to add data from records generated via imitation learning. | LearningExperience.h | |
TSharedRef< FJsonObject > AsJsonConfig
(
const int32 ReplayBufferId |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetActionModifiers
(
const int32 Index |
LearningExperience.h | ||
int32 GetActionModifiersNum() |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetActions
(
const int32 Index |
LearningExperience.h | ||
int32 GetActionsNum() |
LearningExperience.h | ||
const TLearningArrayView< 1, const ECompletionMode > GetEpisodeCompletionModes() |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetEpisodeFinalMemoryStates
(
const int32 Index |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetEpisodeFinalObservations
(
const int32 Index |
LearningExperience.h | ||
const TLearningArrayView< 1, const int32 > GetEpisodeLengths() |
LearningExperience.h | ||
int32 GetEpisodeNum() |
LearningExperience.h | ||
const TLearningArrayView< 1, const int32 > GetEpisodeStarts() |
LearningExperience.h | ||
int32 GetMaxEpisodeNum() |
LearningExperience.h | ||
int32 GetMaxStepNum() |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetMemoryStates
(
const int32 Index |
LearningExperience.h | ||
int32 GetMemoryStatesNum() |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetObservations
(
const int32 Index |
LearningExperience.h | ||
int32 GetObservationsNum() |
LearningExperience.h | ||
const TLearningArrayView< 2, const float > GetRewards
(
const int32 Index |
LearningExperience.h | ||
int32 GetRewardsNum() |
LearningExperience.h | ||
int32 GetStepNum() |
LearningExperience.h | ||
bool HasCompletions() |
LearningExperience.h | ||
bool HasFinalMemoryStates() |
LearningExperience.h | ||
bool HasFinalObservations() |
LearningExperience.h | ||
void Reset() |
Reset the replay buffer. Does not free memory - just resets episode and sample num to zero. | LearningExperience.h | |
void Resize
(
const FEpisodeBuffer& EpisodeBuffer, |
Resizes the replay buffer. | LearningExperience.h |