Module langbrainscore.benchmarks
Expand source code
import typing
from pathlib import Path
from functools import partial
from langbrainscore.benchmarks.pereira2018 import pereira2018_mean_froi
supported_benchmarks: typing.Mapping[str, typing.Callable] = {
"pereira2018_mean_froi_Lang": partial(pereira2018_mean_froi, network="Lang"),
"pereira2018_mean_froi_MD": partial(pereira2018_mean_froi, network="MD"),
"pereira2018_mean_froi": pereira2018_mean_froi,
"pereira2018_ind_voxels": NotImplemented,
}
def load_benchmark(
benchmark_name_or_path: typing.Union[str, Path],
*loading_args,
load_cache=True,
**loading_kwargs,
) -> "langbrainscore.dataset.Dataset":
"""A method that, given a name or a path to a benchmark, loads it and returns
a dataset object.
Args:
benchmark_name_or_path (typing.Union[str, Path]): name of a pre-packaged/officially supported benchmark
or a path to a benchmark with valid formatting
Returns:
langbrainscore.dataset.Dataset: a langbrainscore Dataset object
"""
if benchmark_name_or_path in supported_benchmarks:
loader = supported_benchmarks[benchmark_name_or_path]
return loader(*loading_args, load_cache=load_cache, **loading_kwargs)
raise NotImplementedError(
f"Benchmark identified by `{benchmark_name_or_path}` currently not supported."
)
Sub-modules
langbrainscore.benchmarks.pereira2018
Functions
def load_benchmark(benchmark_name_or_path: Union[str, pathlib.Path], *loading_args, load_cache=True, **loading_kwargs) ‑> langbrainscore.dataset.Dataset
-
A method that, given a name or a path to a benchmark, loads it and returns a dataset object.
Args
benchmark_name_or_path
:typing.Union[str, Path]
- name of a pre-packaged/officially supported benchmark or a path to a benchmark with valid formatting
Returns
langbrainscore.dataset.Dataset
- a langbrainscore Dataset object
Expand source code
def load_benchmark( benchmark_name_or_path: typing.Union[str, Path], *loading_args, load_cache=True, **loading_kwargs, ) -> "langbrainscore.dataset.Dataset": """A method that, given a name or a path to a benchmark, loads it and returns a dataset object. Args: benchmark_name_or_path (typing.Union[str, Path]): name of a pre-packaged/officially supported benchmark or a path to a benchmark with valid formatting Returns: langbrainscore.dataset.Dataset: a langbrainscore Dataset object """ if benchmark_name_or_path in supported_benchmarks: loader = supported_benchmarks[benchmark_name_or_path] return loader(*loading_args, load_cache=load_cache, **loading_kwargs) raise NotImplementedError( f"Benchmark identified by `{benchmark_name_or_path}` currently not supported." )