FMinariCollectionSettings

Configuration struct for Minari dataset collection.

Contains all settings necessary to collect an imitation learning dataset using Minari. This is a Minari-specific implementation of imitation learning settings.

struct FMinariCollectionSettings : public FImitationSettings

Methods

GenerateImitationArgs

virtual void GenerateImitationArgs(FScriptArgBuilder &ArgBuilder) const override

Generate command-line arguments for the Minari collection script.

Parameters

• ArgBuilder (FScriptArgBuilder)

---

FMinariCollectionSettings

virtual ~FMinariCollectionSettings()

Virtual destructor.

Attributes

TEXT

FString DatasetId = TEXT("my-dataset-v0")

Unique identifier for the Minari dataset.

Parameters

• "my-dataset-v0"

---

NumSteps

int32 NumSteps = 1000

Total number of steps to collect for the dataset.

---

Seed

int32 Seed = -1

Random seed for reproducibility.

---

bIncludeAuthor

bool bIncludeAuthor = false

Author name for the dataset metadata.

---

Author

FString Author

Author name for the dataset metadata.

---

bIncludeAuthorEmail

bool bIncludeAuthorEmail = false

Author email for the dataset metadata.

---

AuthorEmail

FString AuthorEmail

Author email for the dataset metadata.

---

bIncludeCodePermalink

bool bIncludeCodePermalink = false

URL to the code or repository used to generate the dataset.

---

CodePermalink

FString CodePermalink

URL to the code or repository used to generate the dataset.

---

bIncludeAlgorithmName

bool bIncludeAlgorithmName = false

Name of the algorithm or policy used to collect the data.

---

AlgorithmName

FString AlgorithmName

Name of the algorithm or policy used to collect the data.

---

bIncludeDescription

bool bIncludeDescription = false

Description of the dataset.

---

Description

FString Description

Description of the dataset.

---

bRecordInfos

bool bRecordInfos = false

Whether to record the info dictionaries in the dataset.

---

bUseCustomDataPath

bool bUseCustomDataPath = false

Directory path where Minari datasets will be stored.

---

DataPath

FDirectoryPath DataPath

Directory path where Minari datasets will be stored.

---

LoggingSettings

FImitationLoggingSettings LoggingSettings

Settings for logging during data collection.

Source: Source/ScholaImitation/Public/ImitationSettings/Minari/MinariCollectionSettings.h