nessie.models.featurizer

Module Contents

Classes

CachedSentenceTransformer
FlairTokenEmbeddingsWrapper
SentenceEmbedder
TfIdfSentenceEmbedder

class nessie.models.featurizer.CachedSentenceTransformer(model_name: str = SBERT_MODEL_NAME, cache_dir: Optional[pathlib.Path] = None)

Bases: SentenceEmbedder

embed(self, sentences: nessie.types.StringArray) → numpy.typing.NDArray[str]

get_dimension(self) → int

class nessie.models.featurizer.FlairTokenEmbeddingsWrapper(embedder: flair.embeddings.TokenEmbeddings, batch_size: int = 8)

embed(self, sentences: nessie.types.RaggedStringArray, flat: bool = False) → Union[numpy.typing.NDArray[float], awkward.Array]

property embedding_dim(self) → int

class nessie.models.featurizer.SentenceEmbedder

abstract embed(self, sentences: nessie.types.StringArray) → numpy.typing.NDArray[str]

eval(self)

abstract get_dimension(self) → int

train(self)

class nessie.models.featurizer.TfIdfSentenceEmbedder

Bases: SentenceEmbedder

embed(self, sentences: nessie.types.StringArray) → numpy.typing.NDArray[str]

eval(self)

get_dimension(self) → int

train(self)