DatasetDiscoveryConfig¶
Module: fast_llm.data.preparation.dataset_discovery.config
Variant of: RunnableConfig — select with type: prepare_dataset_discovery
Variant of: DatasetPreparatorConfig — select with type: dataset_discovery
Inherits from: DatasetPreparatorConfig, RunnableConfig
Fields¶
directory—core-
Type:
PathDefault: (required)Directory to search for datasets recursively
output—core-
Type:
PathDefault: (required)Output path for the generated config YAML file
ignore_paths—optional-
Type: list[
Path] Default:list()List of paths to ignore during dataset discovery (can be absolute or relative to directory)