README.md

Jaseci NLP Package (jac_nlp)

The jac_nlp package contains a collection of state-of-the-art NLP models that can be used to perform various nlp tasks such as named entity recongnition, text summerization, embedding generation, topic extraction etc. following is a list of all the models available in the jac_nlp package.

Installation

Each module can be installed individually or all at once. To install all modules at once.

pip install jac_nlp[all] # Installs all the modules present in the jac_nlp package
pip install jac_nlp[use_enc] # Installs the use_enc module present in the jac_nlp package
pip install jac_nlp[use_qa,ent_ext] # Installs the use_qa and ent_ext modules present in the jac_nlp package

List of Models

Module	Model Type	Model Name	Docs	Type	Status	Description	Resources
use_enc	Text Encoder	USE Encoder	Link	Zero-shot	Ready	Sentence-level embedding pre-trained on general text corpus	Paper
use_qa	Text Encoder	USE QA	Link	Zero-shot	Ready	Sentence-level embedding pre-trained on Q&A data corpus	Paper
fast_enc	Text Encoder	FastText	Link	Training req.	Ready	FastText Text Classifier	Paper
bi_enc	Text Encoder	Bi-encoder	Link	Training req./Zero-shot	Ready	Dual sentence-level encoders	Paper
sbert_sim	Text Encoder	SBert Similarity	Link	Training req./Zero-shot	Ready	SBert Encoders for Sentence Similarity	Paper
ent_ext/ lstm_ner	Named Entity Recognition	Flair NER	Link	Training req.	Ready	Entity extraction using the FLAIR NER framework
tfm_ner	Named Entity Recognition	Transformer NER	Link	Training req.	Ready	Token classification on Transformer models, can be used for NER	Huggingface
cl_summer	Summarization	Summarizer	Link	No Training req.	Ready	Extractive Summarization using Sumy	Doc.
t5_sum	Summarization	Summarizer	Link	No Training req.	Ready	Abstractive Summarization using the T5 Model	Doc., Paper
bart_sum	Summarization	Summarizer	Link	No Training req.	Ready	Abstractive Summarization using the Bart Large Model	Huggingface, Paper
text_seg	Text Processing	Text Segmenter	Link	No Training req.	Experimetal	Topical Change Detection in Documents	Huggingface
topic_ext	Text Analysis	Topic Extraction	Link	No Training req.	Experimetal	Indentifying most relevent topics for given set of documents
sentiment	Text Analysis	Sentiment Analysis	Link	No Training req.	Experimetal	Determining the overall sentiment expressed is text as positive, negative, or neutral

Usage

To load the jac_nlp.use_enc package into jaseci in local environment, run the following command in the jsctl console.

jsctl > actions load module jac_nlp.use_enc