Module: tfm.nlp.models

Models are combinations of tf.keras layers and models that can be trained.

Several pre-built canned models are provided to train encoder networks. These models are intended as both convenience functions and canonical examples.

Classes

class BertClassifier: Classifier model based on a BERT-style transformer-based encoder.

class BertPretrainer: BERT pretraining model.

class BertPretrainerV2: BERT pretraining model V2.

class BertSpanLabeler: Span labeler model based on a BERT-style transformer-based encoder.

class BertTokenClassifier: Token classifier model based on a BERT-style transformer-based encoder.

class DualEncoder: A dual encoder model based on a transformer-based encoder.

class ElectraPretrainer: ELECTRA network training model.

class Seq2SeqTransformer: Transformer model with Keras.

class T5Transformer: Transformer Encoder+Decoder for sequence to sequence.

class T5TransformerParams: Transformer parameters.

class TransformerDecoder: Transformer decoder.

class TransformerEncoder: Transformer encoder.

class XLNetClassifier: Classifier model based on XLNet.

class XLNetPretrainer: XLNet-based pretrainer.

class XLNetSpanLabeler: Span labeler model based on XLNet.

Functions

attention_initializer(...): Initializer for attention layers in Seq2SeqTransformer.

EOS_ID 1