Sentiment Analysis with Movies Reviews

This repo is a sentiment analysis classifier built with Transformer based models

• The whole experiment environment is tested and built on Google Colab.
• This classifier is built with Transformer based models, LSTM and other NLP classifiers.

Requirements

• This project is allowed to run on Google Colab with single GPU.
  If you prefer to work on local environment, this repo is able to run using Jupyter Notebook.
• Since BERT model is developed under Tensorflow 1.x, more modification to get compatible with Tensorflow 2.x is already added to codes.

Datasets

• IMDB Reviews
• Classify the sentiment of sentences from the Rotten Tomatoes dataset

Run the classifier

To run the classifier, follow the instructions below...

• If running on local environment, import BERT_Demo to Jupyter Notebook.
• If running on Google Colab, import BERT_Demo directly.