Model packageΒΆ

This package contains all the model implementations in the repository, and some extra logic. See the documentations for the individual models below for more information:

  • nlp_uncertainty_zoo.models.bayesian_lstm: Implementation of the Bayesian LSTM

  • nlp_uncertainty_zoo.models.bert: Implementation of a BERT wrapper class.

  • nlp_uncertainty_zoo.models.ddu_transformer: Implementation of a Deep Deterministic Uncertaity Transformer and BERT.

  • nlp_uncertainty_zoo.models.dpp_transformer: Implementation of a Determinantal Point-Process Transformer and BERT.

  • nlp_uncertainty_zoo.models.lstm: Implementation of Long-short term memory RNN.

  • nlp_uncertainty_zoo.models.lstm_ensemble: Implementation of an ensemble of LSTMs.

  • nlp_uncertainty_zoo.models.model: Explanation of the abstract Model and Module classes used in the repository.

  • nlp_uncertainty_zoo.models.sngp_transformer: Implementation of the Spectral-normalized Gaussian Process Transformer and BERT.

  • nlp_uncertainty_zoo.models.spectral: Implementation of superclasses using spectral normalization (used for nlp_uncertainty_zoo.models.ddu_transformer and nlp_uncertainty_zoo.models.sngp_transformer)

  • nlp_uncertainty_zoo.models.st_tau_lstm: Implementation of the ST-tau LSTM.

  • nlp_uncertainty_zoo.models.transformer: Implementation of a basic transformer.

  • nlp_uncertainty_zoo.models.variational_lstm: Implementation of a Variational LSTM using MC Dropout.

  • nlp_uncertainty_zoo.models.variational_transformer: Implementation of a Variational Transformer and BERT.