nlpipes 0.1.0

Text Classification with TensorFlow

NLPipes

A Tensorflow Library for Text Classification

Table of Contents

1. Overview
   • Built With
2. How to Use
• Mono Label task

Overview

NLPipes is a Tensorflow library that provides an easy way to train and use deep-learning based models for a diversity of text classification tasks.

NLPipes supports the following tasks:

• Mono-label classification: Assign one label to each text (e.g. positive, neutral, negative).
• Multi-label classification: Assign one or more label to each text from a list of possible classes (e.g., Product Category, Product Quality, Delivery Time, and Price)
• Class-label classification: Assign one label from a list of possible labels for each of a list of classes. A typical use case is aspect based sentiment analysis where one want to detect each aspect mentionned in a review along his assocated sentiment polarity.

Built with

NLPipes is built with TensorFlow and HuggingFace Transformers:

• TensorFlow: An end-to-end open source deep learning framework
• Transformers: An general-purpose open-sources library for transformers-based architectures

How To Use

NLPipes provides high level abstractions to save having to deal with the complexity of deep-learning based Natural Language Understanding algorithms.

Mono-label classification

Give NLPipes a single label for each text as a target.

from nlpipes import Model

model = Model(task = 'mono-label-classification',
              name_or_path = 'bert-base-uncased')

reviews   = ["text", "text", "text", ...]
sentiments = ["positive", "negative", "neutral", ...]

test_reviews = ["text", "text", "text", ...]
test_sentiments = ["positive", "negative", "neutral", ...]

new_reviews = ["text", "text", "text", ...]

model.train(reviews, sentiments)

evaluation = model.evaluate(test_reviews, test_sentiments)

predictions = model.predict(new_reviews)
 
model.save('./sentiment_detection_model')