BERT

SentimentAI: BERT Movie Review Sentiment Analysis

A premium, end-to-end sentiment analysis application using a fine-tuned BERT model on the IMDB movie review dataset.

Features

• BERT Backend: fine-tuned bert-base-uncased for high-accuracy sentiment classification.
• Fast Training: Configurable script to train on subsets or the full dataset.
• Modern UI: A stunning glassmorphism dashboard for real-time inference.
• REST API: Flask-based API for easy integration.

Project Structure

• prepare_data.py: Downloads and tokenizes the IMDB dataset.
• train_bert.py: Fine-tunes the BERT model.
• predict.py: Inference class for sentiment prediction.
• app/: Contains the web dashboard and Flask backend.
• requirements.txt: Python dependencies.

Setup Instructions

1. Install Dependencies

pip install -r requirements.txt

2. Train the Model

Run the training script to fine-tune BERT on the IMDB dataset.

python train_bert.py

Note: By default, it trains on a small subset (1000 samples) for speed. You can increase subset_size in the script for better accuracy.

3. Run the Web App

1. Start the Flask backend:

python app/backend.py

1. Open app/index.html in your browser.

Technologies Used

• Python: Core logic and model training.
• Hugging Face Transformers: BERT model implementation.
• PyTorch: Deep learning framework.
• Flask: Web API.
• Vanilla HTML/CSS/JS: Premium frontend with glassmorphism design.