notice

This is unreleased documentation for Rasa Open Source Documentation Master/Unreleased version.
For the latest released documentation, see the latest version (2.8.x).

Version: Master/Unreleased

rasa.nlu.featurizers.dense_featurizer.dense_featurizer

DenseFeaturizer2 Objects

class DenseFeaturizer2(Featurizer2[np.ndarray], ABC)

Base class for all dense featurizers.

aggregate_sequence_features

@staticmethod
def aggregate_sequence_features(dense_sequence_features: np.ndarray, pooling_operation: Text, only_non_zero_vectors: bool = True) -> np.ndarray

Aggregates the non-zero vectors of a dense sequence feature matrix.

Arguments:

  • dense_sequence_features - a 2-dimensional matrix where the first dimension is the sequence dimension over which we want to aggregate of shape [seq_len, feat_dim]
  • pooling_operation - either max pooling or average pooling
  • only_non_zero_vectors - determines whether the aggregation is done over non-zero vectors only

Returns:

a matrix of shape [1, feat_dim]