A raw input example consisting of one or two segments of text and a label

A dataset of tensors that uses a dictionary for key-value mappings

Configuration for evaluating a model.

A set of numeric features obtained from an :class:`InputExample`

A list of logits obtained from a finetuned PET model

Configuration for training a model.

A wrapper around a Transformer-based language model.

A configuration for a :class:`TransformerModelWrapper`.

Processor for the BoolQ data set.

Processor for the CB data set.

Processor for the COPA data set.

A custom task helper for the COPA dataset.

Abstract class that provides methods for loading train/dev32/dev/test/unlabeled examples for a given task.

Preprocessor for models pretrained using a masked language modeling objective (e.g., BERT).

Processor for the MultiRC data set.

A custom task helper for the MultiRC dataset.

Example for a data processor.

Example for a pattern-verbalizer pair (PVP).

A set of numeric input features for a model pretrained with a permuted language modeling objective.

This class contains functions to apply patterns and verbalizers as required by PET. Each task requires its own custom implementation of a PVP

Abstract class for a PET configuration that can be saved to and loaded from a json file.

A preprocessor that transforms an :class:`InputExample` into a :class:`InputFeatures` object so that it can be processed by the model being u

Processor for the ReCoRD data set.

Processor for the RTE data set.

A helper class that provides custom training and evaluation methods for tasks that do not fit in PETs default schema, for example because the

Processor for the WiC data set.

Processor for the WSC data set.

A custom task helper for the Wsc dataset.