DLTransformer

class lightautoml.text.dl_transformers.DLTransformer(model, model_params, dataset, dataset_params, loader_params, device='cuda', random_state=42, embedding_model=None, embedding_model_params=None, multigpu=False, verbose=False)[source]

Bases: sklearn.base.TransformerMixin

Deep Learning based sentence embeddings.

__init__(model, model_params, dataset, dataset_params, loader_params, device='cuda', random_state=42, embedding_model=None, embedding_model_params=None, multigpu=False, verbose=False)[source]

Class to compute sentence embeddings from words embeddings.

Parameters

  • model – Torch model for aggregation word embeddings into sentence embedding.

  • model_params – Dict with model parameters.

  • dataset – Torch dataset.

  • dataset_params – Dict with dataset params.

  • loader_params – Dict with params for torch dataloader.

  • device – String with torch device type or device ids. I.e: ‘0,2’.

  • random_state – Determines random number generation.

  • embedding_model – Torch word embedding model, if dataset do not return embeddings.

  • embedding_model_params – Dict with embedding model params.

  • multigpu – Use data parallel for multiple GPU.

  • verbose – Show tqdm progress bar.

get_name()[source]

Module name.

Return type

str

Returns

String with module name.

get_out_shape()[source]

Output shape.

Return type

int

Returns

Int with module output shape.