euroeval.benchmarker

Class that benchmarks language models.

Classes

Benchmarker — Benchmarking all the language models.

Functions

model_has_been_benchmarked — Checks whether a model has already been benchmarked on a dataset.
adjust_logging_level — Adjust the logging level based on verbosity.
clear_model_cache_fn — Clear the model cache.
prepare_dataset_configs — Prepare the dataset configuration(s) to be benchmarked.
initial_logging — Initial logging at the start of the benchmarking process.

source class Benchmarker(progress_bar: bool = True, save_results: bool = True, task: str | list[str] | None = None, dataset: list[str] | str | None = None, language: str | list[str] = 'all', model_language: str | list[str] | None = None, dataset_language: str | list[str] | None = None, device: Device | None = None, batch_size: int = 32, raise_errors: bool = False, cache_dir: str = '.euroeval_cache', api_key: str | None = None, force: bool = False, verbose: bool = False, trust_remote_code: bool = False, clear_model_cache: bool = False, evaluate_test_split: bool = False, few_shot: bool = True, num_iterations: int = 10, api_base: str | None = None, api_version: str | None = None, debug: bool = False, run_with_cli: bool = False, only_allow_safetensors: bool = False)

Benchmarking all the language models.

Initialise the benchmarker.

Attributes

benchmark_config_default_params — The default parameters for the benchmark configuration.
benchmark_config — The benchmark configuration.
force — Whether to force evaluations of models, even if they have been benchmarked already.
results_path — The path to the results file.
benchmark_results : list[BenchmarkResult] — The benchmark results.

Parameters

progress_bar : bool — Whether progress bars should be shown. Defaults to True.
save_results : bool — Whether to save the benchmark results to 'euroeval_benchmark_results.jsonl'. Defaults to True.
task : str | list[str] | None — The tasks benchmark the model(s) on. Mutually exclusive with dataset. If both task and dataset are None then all datasets will be benchmarked.
dataset : list[str] | str | None — The datasets to benchmark on. Mutually exclusive with task. If both task and dataset are None then all datasets will be benchmarked.
language : str | list[str] — The language codes of the languages to include, both for models and datasets. Set this to 'all' if all languages should be considered. Defaults to "all".
model_language : str | list[str] | None — The language codes of the languages to include for models. If specified then this overrides the language parameter for model languages. Defaults to None.
dataset_language : str | list[str] | None — The language codes of the languages to include for datasets. If specified then this overrides the language parameter for dataset languages. Defaults to None.
device : Device | None — The device to use for benchmarking. Defaults to None.
batch_size : int — The batch size to use. Defaults to 32.
raise_errors : bool — Whether to raise errors instead of skipping the model evaluation. Defaults to False.
cache_dir : str — Directory to store cached models. Defaults to '.euroeval_cache'.
api_key : str | None — The API key to use for a given inference API.
force : bool — Whether to force evaluations of models, even if they have been benchmarked already. Defaults to False.
verbose : bool — Whether to output additional output. This is automatically set if debug is True. Defaults to False.
trust_remote_code : bool — Whether to trust remote code when loading models. Defaults to False.
clear_model_cache : bool — Whether to clear the model cache after benchmarking each model. Defaults to False.
evaluate_test_split : bool — Whether to evaluate the test split of the datasets. Defaults to False.
few_shot : bool — Whether to only evaluate the model using few-shot evaluation. Only relevant if the model is generative. Defaults to True.
num_iterations : int — The number of times each model should be evaluated. This is only meant to be used for power users, and scores will not be allowed on the leaderboards if this is changed. Defaults to 10.
api_base : str | None — The base URL for a given inference API. Only relevant if model refers to a model on an inference API. Defaults to None.
api_version : str | None — The version of the API to use. Defaults to None.
debug : bool — Whether to output debug information. Defaults to False.
run_with_cli : bool — Whether the benchmarker is being run from the command-line interface. Defaults to False.
only_allow_safetensors : bool — Whether to only allow models that use the safetensors format. Defaults to False.

Raises

ValueError — If both task and dataset are specified.

Methods

benchmark — Benchmarks models on datasets.

source property Benchmarker.benchmark_results: list[BenchmarkResult]

The benchmark results.

Benchmarks models on datasets.

Parameters

model : list[str] | str — The full Hugging Face Hub path(s) to the pretrained transformer model. The specific model version to use can be added after the suffix '@': "model@v1.0.0". It can be a branch name, a tag name, or a commit id, and defaults to the latest version if not specified.
task : str | list[str] | None — The tasks benchmark the model(s) on. Mutually exclusive with dataset. If both task and dataset are None then all datasets will be benchmarked. Defaults to None.
dataset : list[str] | str | None — The datasets to benchmark on. Mutually exclusive with task. If both task and dataset are None then all datasets will be benchmarked. Defaults to None.
progress_bar : bool | None — Whether progress bars should be shown. Defaults to the value specified when initialising the benchmarker.
save_results : bool | None — Whether to save the benchmark results to 'euroeval_benchmark_results.jsonl'. Defaults to the value specified when initialising the benchmarker.
language : str | list[str] | None — The language codes of the languages to include, both for models and datasets. Here 'no' means both Bokmål (nb) and Nynorsk (nn). Set this to 'all' if all languages should be considered. Defaults to the value specified when initialising the benchmarker.
model_language : str | list[str] | None — The language codes of the languages to include for models. If specified then this overrides the language parameter for model languages. Defaults to the value specified when initialising the benchmarker.
dataset_language : str | list[str] | None — The language codes of the languages to include for datasets. If specified then this overrides the language parameter for dataset languages. Defaults to the value specified when initialising the benchmarker.
device : Device | None — The device to use for benchmarking. Defaults to the value specified when initialising the benchmarker.
batch_size : int | None — The batch size to use. Defaults to the value specified when initialising the benchmarker.
raise_errors : bool | None — Whether to raise errors instead of skipping the model evaluation.
cache_dir : str | None — Directory to store cached models. Defaults to the value specified when initialising the benchmarker.
api_key : str | None — The API key to use for a given inference server. Defaults to the value specified when initialising the benchmarker.
force : bool | None — Whether to force evaluations of models, even if they have been benchmarked already. Defaults to the value specified when initialising the benchmarker.
verbose : bool | None — Whether to output additional output. Defaults to the value specified when initialising the benchmarker.
trust_remote_code : bool | None — Whether to trust remote code when loading models. Defaults to the value specified when initialising the benchmarker.
clear_model_cache : bool | None — Whether to clear the model cache after benchmarking each model. Defaults to the value specified when initialising the benchmarker.
evaluate_test_split : bool | None — Whether to evaluate the test split of the datasets. Defaults to the value specified when initialising the benchmarker.
few_shot : bool | None — Whether to only evaluate the model using few-shot evaluation. Only relevant if the model is generative. Defaults to the value specified when initialising the benchmarker.
num_iterations : int | None — The number of times each model should be evaluated. This is only meant to be used for power users, and scores will not be allowed on the leaderboards if this is changed. Defaults to the value specified when initialising the benchmarker.
only_allow_safetensors : bool | None — Whether to only allow models that use the safetensors format. Defaults to the value specified when initialising the benchmarker.

Returns

list[BenchmarkResult] — A list of benchmark results.

Raises

ValueError — If both task and dataset are specified.
benchmark_output_or_err
e

source model_has_been_benchmarked(model_id: str, dataset: str, few_shot: bool, validation_split: bool, benchmark_results: list[BenchmarkResult]) → bool

Checks whether a model has already been benchmarked on a dataset.

Parameters

model_id : str — The model ID.
dataset : str — The dataset.
few_shot : bool — Whether the model was evaluated using few-shot evaluation.
validation_split : bool — Whether the model was evaluated on the validation split.
benchmark_results : list[BenchmarkResult] — The benchmark results.

Returns

bool — Whether the model has already been evaluated on the dataset.

source adjust_logging_level(verbose: bool, ignore_testing: bool = False) → int

Adjust the logging level based on verbosity.

Parameters

verbose : bool — Whether to output additional output.
ignore_testing : bool — Whether to ignore the testing flag.

Returns

int — The logging level that was set.

source clear_model_cache_fn(cache_dir: str) → None

Clear the model cache.

Note that this will not remove the stored completions.

Parameters

cache_dir : str — The path to the cache directory.

source prepare_dataset_configs(dataset_names: list[str]) → list['DatasetConfig']

Prepare the dataset configuration(s) to be benchmarked.

Parameters

dataset_names : list[str] — The dataset names to benchmark.

Returns

list['DatasetConfig'] — The prepared list of model IDs.

source initial_logging(model_config: ModelConfig, dataset_config: DatasetConfig, benchmark_config: BenchmarkConfig) → None

Initial logging at the start of the benchmarking process.

Parameters

model_config : ModelConfig — The configuration of the model we are evaluating.
dataset_config : DatasetConfig — The configuration of the dataset we are evaluating on.
benchmark_config : BenchmarkConfig — The general benchmark configuration.