Long Short Term Memory

class sentivi.classifier.LSTMClassifier(num_labels: int, embedding_size: Optional[int] = None, max_length: Optional[int] = None, device: Optional[str] = 'cpu', num_epochs: Optional[int] = 10, learning_rate: Optional[float] = 0.001, batch_size: Optional[int] = 2, shuffle: Optional[bool] = True, random_state: Optional[int] = 101, hidden_size: Optional[int] = 512, hidden_layers: Optional[int] = 2, bidirectional: Optional[bool] = False, attention: Optional[bool] = True, *args, **kwargs)
__init__(num_labels: int, embedding_size: Optional[int] = None, max_length: Optional[int] = None, device: Optional[str] = 'cpu', num_epochs: Optional[int] = 10, learning_rate: Optional[float] = 0.001, batch_size: Optional[int] = 2, shuffle: Optional[bool] = True, random_state: Optional[int] = 101, hidden_size: Optional[int] = 512, hidden_layers: Optional[int] = 2, bidirectional: Optional[bool] = False, attention: Optional[bool] = True, *args, **kwargs)

Initialize LSTMClassifier

Parameters

  • num_labels – number of polarities

  • embedding_size – input embeddings’ size

  • max_length – maximum length of input text

  • device – training device

  • num_epochs – maximum number of epochs

  • learning_rate – model learning rate

  • batch_size – training batch size

  • shuffle – whether DataLoader is shuffle or not

  • random_state – random.seed number

  • hidden_size – Long Short Term Memory hidden size

  • bidirectional – whether to use BiLSTM or not

  • args – arbitrary arguments

  • kwargs – arbitrary keyword arguments

forward(data, *args, **kwargs)

Training and evaluating methods

Parameters

  • data – TextEncoder output

  • args – arbitrary arguments

  • kwargs – arbitrary keyword arguments

Returns

training results

predict(X, *args, **kwargs)

Predict polarity with given sentences

Parameters

  • X – TextEncoder.predict output

  • args – arbitrary arguments

  • kwargs – arbitrary keyword arguments

Returns

list of numeric polarities

Return type

list