Skip to content

euroeval.data_loading

source module euroeval.data_loading

Functions related to the loading of the data.

Functions

source load_data(rng: Generator, dataset_config: DatasetConfig, benchmark_config: BenchmarkConfig) → list[DatasetDict]

Load the raw bootstrapped datasets.

Parameters

  • rng : Generator — The random number generator to use.

  • dataset_config : DatasetConfig — The configuration for the dataset.

  • benchmark_config : BenchmarkConfig — The configuration for the benchmark.

Returns

  • list[DatasetDict] — A list of bootstrapped datasets, one for each iteration.

Raises

  • InvalidBenchmark — If the dataset cannot be loaded.

  • HuggingFaceHubDown — If the Hugging Face Hub is down.

source load_raw_data(dataset_config: DatasetConfig, cache_dir: str) → DatasetDict

Load the raw dataset.

Parameters

  • dataset_config : DatasetConfig — The configuration for the dataset.

  • cache_dir : str — The directory to cache the dataset.

Returns

  • DatasetDict — The dataset.

Raises