logitorch.datasets.proof_qa.fld_dataset

Attributes

FLD_SUB_DATASETS
FLD_TASKS

Classes

FLDDataset

Helper class that provides a standard way to create an ABC using

Module Contents

class logitorch.datasets.proof_qa.fld_dataset.FLDDataset(dataset_name: str, split_set: str, task: str, max_samples: int | None = None)

Bases: logitorch.datasets.base.AbstractProofQADataset

Helper class that provides a standard way to create an ABC using inheritance.

Initializes an instance of the FLDDataset class.

Args:

dataset_name (str): The name of the dataset. Must be one of the FLD sub-datasets: “FLD.v2” or “FLD_star.v2”. split_set (str): The split set of the dataset. Must be one of the predefined split sets. task (str): The task to perform on the dataset. Must be “proof_generation_all”. max_samples (Optional[int], optional): The maximum number of samples to load from the dataset. Defaults to None.

Raises:

DatasetNameError: If the dataset_name is not one of the FLD sub-datasets. SplitSetError: If the split_set is not one of the predefined split sets. TaskError: If the task is not “proof_generation_all”.

logitorch.datasets.proof_qa.fld_dataset.FLD_SUB_DATASETS = ['FLD.v2', 'FLD_star.v2']

logitorch.datasets.proof_qa.fld_dataset.FLD_TASKS = ['proof_generation_all']