Skip to content

Latest commit

 

History

History
28 lines (18 loc) · 1.51 KB

README.md

File metadata and controls

28 lines (18 loc) · 1.51 KB

Movie Reviews Summarization and Sentiment Analysis

Demo

Try out the application at Heroku: https://movie-reviews-687da07d5dc8.herokuapp.com/

Screenshot:

How it works

  1. Searches for Rotten Tomatoes ${query} on google and retrieves the URL of the top result.
  2. Scrapes the Rotten Tomatoes website for the top critic reviews using Selenium.
  3. Uses Hugging Face Inference API to generate the summary and perform the sentiment analysis of the movie reviews.

Summarization Model

Model: https://huggingface.co/abhiramd22/t5-base-finetuned-to-summarize-movie-reviews

This is based on t5-base model, fine-tuned using a custom-built movie reviews dataset.

Sentiment Analysis Model

Model: https://huggingface.co/abhiramd22/finetuning-sentiment-model-mpnet-imdb

This is based on all-mpnet-base-v2 model, fine-tuned using the IMDB movie reviews dataset.

Jupyter Notebooks

The Jupyter Notebooks used to train & fine-tune the classification and summarization models used in the web application can be found at /Jupyter_Notebooks

The Input and Intermediate data used in these notebooks can be found at https://github.com/Abhiram-DLVSS/Movie-Reviews-Analysis-Data