Sentiment-analysis-using-DistillBERT-and-ktrain

Built a text classification model that can be used for predicting user reviews using DistillBert transformer.
Dataset used contains two classes.
Positive and negative sentiment are the two classes.
Used ktrain for preprocessing and training the text data.
Trained and saved the model with prediction result of 94% accuracy.
Deployed the model as web aplication using flask framework.

Code and Resources Used

Python Version: 3.7

Tensorflow version: 2.1.0

Packages: pandas, numpy, tensorflow, ktrain, flask

ktrain implementation reference : Laxmi kant

Dataset Structure

Two folders

Train: 25000 reviews
Test: 25000 reviews

Model deployment using Flask

Using the flask framework, the model was deployed in web application.
Model can be deployed using 'flask_DistillBERT_ktrain.py'

Positive sentiment prediction example

Negative sentiment prediction example