pine.quantitative_evaluation.language_modeling package

Submodules

pine.quantitative_evaluation.language_modeling.data module

class pine.quantitative_evaluation.language_modeling.data.Corpus(dataset: Dict[str, pathlib.Path], dictionary: Optional[pine.quantitative_evaluation.language_modeling.data.Dictionary] = None)

Bases: object

tokenize(subset: str) → torch.Tensor

tokenize_all()

class pine.quantitative_evaluation.language_modeling.data.Dictionary

Bases: object

add_word(word)

pine.quantitative_evaluation.language_modeling.data.read_text_file(path: pathlib.Path) → Iterable[str]

pine.quantitative_evaluation.language_modeling.data.simple_preprocess(text: str) → List[str]

pine.quantitative_evaluation.language_modeling.language_modeling module

class pine.quantitative_evaluation.language_modeling.language_modeling.Result(language_model: pine.language_model.LanguageModel, result: Tuple[Tuple[float, float], Sequence[Tuple[Sequence[Tuple[float, float]], Tuple[float, float], float]]])

Bases: object

The results of a language modeling task for a log-bilinear language model.

Parameters

• language_model (LanguageModel) – A log-bilinear language model.

• result (RawResult) – Results of a language modeling task.

Variables

• language_model (LanguageModel) – A log-bilinear language model.

• result (RawResult) – Results of a language modeling task.

plot(*args: Any, **kwargs: Any) → Figure

pine.quantitative_evaluation.language_modeling.language_modeling.evaluate(dataset_paths: Dict[str, pathlib.Path], language_model: pine.language_model.LanguageModel) → pine.quantitative_evaluation.language_modeling.language_modeling.Result

pine.quantitative_evaluation.language_modeling.language_modeling.get_dataset_paths(language: str, dataset_dir: pathlib.Path) → Dict[str, pathlib.Path]

pine.quantitative_evaluation.language_modeling.model module

class pine.quantitative_evaluation.language_modeling.model.PreinitializedRNNModel(vectors: gensim.models.keyedvectors.KeyedVectors, dataset: Dict[str, pathlib.Path])

Bases: pine.quantitative_evaluation.language_modeling.model.RNNModel

synchronize_mapping(dictionary: pine.quantitative_evaluation.language_modeling.data.Dictionary) → pine.quantitative_evaluation.language_modeling.data.Dictionary

class pine.quantitative_evaluation.language_modeling.model.RNNModel(ntoken: int)

Bases: torch.nn.modules.module.Module

forward(input: torch.Tensor, hidden: torch.Tensor) → torch.Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

init_hidden(bsz: int) → Tuple[torch.Tensor, torch.Tensor]

init_weights()

pine.quantitative_evaluation.language_modeling.training module

pine.quantitative_evaluation.language_modeling.training.get_batch(source: torch.Tensor, i: int) → torch.Tensor

pine.quantitative_evaluation.language_modeling.training.repackage_hidden(h: Union[torch.Tensor, Iterable[NestedTensor]]) → Union[torch.Tensor, Iterable[NestedTensor]]

pine.quantitative_evaluation.language_modeling.training.train_and_evaluate(dataset: Dict[str, pathlib.Path], language_model: pine.language_model.LanguageModel) → Tuple[Tuple[float, float], Sequence[Tuple[Sequence[Tuple[float, float]], Tuple[float, float], float]]]

pine.quantitative_evaluation.language_modeling.view module

pine.quantitative_evaluation.language_modeling.view.plot_language_modeling_results(*language_models: pine.language_model.LanguageModel, kind: Optional[str] = None, subset: Optional[str] = None) → matplotlib.figure.Figure

Module contents

class pine.quantitative_evaluation.language_modeling.Result(language_model: pine.language_model.LanguageModel, result: Tuple[Tuple[float, float], Sequence[Tuple[Sequence[Tuple[float, float]], Tuple[float, float], float]]])

Bases: object

The results of a language modeling task for a log-bilinear language model.

Parameters

• language_model (LanguageModel) – A log-bilinear language model.

• result (RawResult) – Results of a language modeling task.

Variables

• language_model (LanguageModel) – A log-bilinear language model.

• result (RawResult) – Results of a language modeling task.

plot(*args: Any, **kwargs: Any) → Figure

pine.quantitative_evaluation.language_modeling.evaluate(dataset_paths: Dict[str, pathlib.Path], language_model: pine.language_model.LanguageModel) → pine.quantitative_evaluation.language_modeling.language_modeling.Result

pine.quantitative_evaluation.language_modeling.get_dataset_paths(language: str, dataset_dir: pathlib.Path) → Dict[str, pathlib.Path]

pine.quantitative_evaluation.language_modeling.plot_language_modeling_results(*language_models: pine.language_model.LanguageModel, kind: Optional[str] = None, subset: Optional[str] = None) → matplotlib.figure.Figure