Utils

The utils package combines certain utility function used within the repository. These include the following:

• nlp_uncertainty_zoo.utils.custom_types: Custom types defined for Python typing annotations.

• nlp_uncertainty_zoo.utils.data: Classes that are used to build batched datasets from raw files.

• nlp_uncertainty_zoo.utils.metrics: Implementation of general-purpose uncertainty metrics.

• nlp_uncertainty_zoo.utils.samplers: Implementation of samplers that produce representative sub-samples of datasets.

• nlp_uncertainty_zoo.utils.task_eval: Functions that are used to evaluate the task performance of models.

• nlp_uncertainty_zoo.utils.uncertainty_eval: Functions that are used to evaluate the calibration and uncertainty quality of models.