API

This API documentation covers each model within NLP Architect. Most modules have a corresponding user guide section that introduces the main concepts. See this API for specific function definitions.

nlp_architect.models

Model classes stores a list of layers describing the model. Methods are provided to train the model weights, perform inference, and save/load the model.

nlp_architect.models.bist_parser.BISTModel BIST parser model class.
nlp_architect.models.chunker.SequenceChunker A sequence Chunker model written in Tensorflow (and Keras) based SequenceTagger model.
nlp_architect.models.intent_extraction.Seq2SeqIntentModel Encoder Decoder Deep LSTM Tagger Model (using tf.keras)
nlp_architect.models.intent_extraction.MultiTaskIntentModel Multi-Task Intent and Slot tagging model (using tf.keras)
nlp_architect.models.matchlstm_ansptr.MatchLSTMAnswerPointer Defines end to end MatchLSTM and Answer_Pointer network for Reading Comprehension
nlp_architect.models.memn2n_dialogue.MemN2N_Dialog End-To-End Memory Network.
nlp_architect.models.most_common_word_sense.MostCommonWordSense
nlp_architect.models.ner_crf.NERCRF Bi-LSTM NER model with CRF classification layer (tf.keras model)
nlp_architect.models.np2vec.NP2vec Initialize the np2vec model, train it, save it and load it.
nlp_architect.models.np_semantic_segmentation.NpSemanticSegClassifier NP Semantic Segmentation classifier model (based on tf.Keras framework).
nlp_architect.models.temporal_convolutional_network.TCN This class defines core TCN architecture.
nlp_architect.models.crossling_emb.WordTranslator Main network which does cross-lingual embeddings training
nlp_architect.models.cross_doc_sieves
nlp_architect.models.cross_doc_coref.sieves_config.EventSievesConfiguration
nlp_architect.models.cross_doc_coref.sieves_config.EntitySievesConfiguration
nlp_architect.models.cross_doc_coref.sieves_resource.SievesResources
nlp_architect.models.gnmt_model.GNMTModel Sequence-to-sequence dynamic model with GNMT attention architecture with sparsity policy support.

nlp_architect.data

Dataset implementations and data loaders (check deep learning framework compatibility of dataset/loader in documentation)

nlp_architect.data.amazon_reviews.Amazon_Reviews Take the *.json file of Amazon reviews as downloaded from http://jmcauley.ucsd.edu/data/amazon/ Then does data cleaning and balancing, as well as transforms the reviews 1-5 to a sentiment
nlp_architect.data.babi_dialog.BABI_Dialog This class loads in the Facebook bAbI goal oriented dialog dataset and vectorizes them into user utterances, bot utterances, and answers.
nlp_architect.data.conll.ConllEntry
nlp_architect.data.intent_datasets.IntentDataset Intent extraction dataset base class
nlp_architect.data.intent_datasets.TabularIntentDataset Tabular Intent/Slot tags dataset loader.
nlp_architect.data.intent_datasets.SNIPS SNIPS dataset class
nlp_architect.data.ptb.PTBDataLoader Class that defines data loader
nlp_architect.data.sequential_tagging.CONLL2000 CONLL 2000 POS/chunking task data set (numpy)
nlp_architect.data.sequential_tagging.SequentialTaggingDataset Sequential tagging dataset loader.
nlp_architect.data.fasttext_emb.FastTextEmb Downloads FastText Embeddings for a given language to the given path.
nlp_architect.data.cdc_resources.relations.computed_relation_extraction.ComputedRelationExtraction Extract Relation between two mentions according to computation and rule based algorithms
nlp_architect.data.cdc_resources.relations.referent_dict_relation_extraction.ReferentDictRelationExtraction
nlp_architect.data.cdc_resources.relations.verbocean_relation_extraction.VerboceanRelationExtraction
nlp_architect.data.cdc_resources.relations.wikipedia_relation_extraction.WikipediaRelationExtraction
nlp_architect.data.cdc_resources.relations.within_doc_coref_extraction.WithinDocCoref
nlp_architect.data.cdc_resources.relations.word_embedding_relation_extraction.WordEmbeddingRelationExtraction
nlp_architect.data.cdc_resources.relations.wordnet_relation_extraction.WordnetRelationExtraction
nlp_architect.data.cdc_resources.relations.relation_types_enums.RelationType An enumeration.

nlp_architect.pipelines

NLP pipelines modules using NLP Architect models

nlp_architect.pipelines.spacy_bist.SpacyBISTParser Main class which handles parsing with Spacy-BIST parser.
nlp_architect.pipelines.spacy_np_annotator.NPAnnotator Spacy based NP annotator - uses models.SequenceChunker model for annotation
nlp_architect.pipelines.spacy_np_annotator.SpacyNPAnnotator Simple Spacy pipe with NP extraction annotations

nlp_architect.contrib

In addition to imported layers, the library also contains its own set of network layers and additions. These are currently stored in the various models or related to which DL frameworks it was based on.

nlp_architect.contrib.tensorflow.python.keras.layers.crf.CRF Conditional Random Field layer (tf.keras) CRF can be used as the last layer in a network (as a classifier).
nlp_architect.contrib.tensorflow.python.keras.utils.layer_utils.save_model Save a model to a file (tf.keras models only) The method save the model topology, as given as a :param model: model object :param topology: a dictionary of topology elements and their values :type topology: dict :param filepath: path to save model :type filepath: str
nlp_architect.contrib.tensorflow.python.keras.utils.layer_utils.load_model Load a model (tf.keras) from disk, create topology from loaded values and load weights.
nlp_architect.contrib.tensorflow.python.keras.callbacks.ConllCallback A Tensorflow(Keras) Conlleval evaluator.

nlp_architect.common

Common types of data structures used by NLP models

nlp_architect.common.core_nlp_doc.CoreNLPDoc Object for core-components (POS, Dependency Relations, etc).
nlp_architect.common.high_level_doc.HighLevelDoc object for annotation documents
nlp_architect.common.cdc.mention_data.MentionDataLight
nlp_architect.common.cdc.mention_data.MentionData