Sentiment Analysis with Hugging Face and Streamlit

Overview

This project demonstrates an end-to-end sentiment analysis application using Hugging Face's transformers library and Streamlit. It allows users to train a sentiment analysis model on the IMDB dataset, evaluate its performance, and make predictions on new text inputs through an interactive web interface.

Features

• Data Loading: Utilizes Hugging Face's datasets library to load the IMDB dataset without local downloads.
• Model Training: Fine-tunes a pre-trained DistilBERT model on the IMDB dataset for sentiment analysis.
• Model Evaluation: Provides performance metrics including accuracy, precision, recall, and F1 score.
• Sentiment Prediction: Allows users to input text and receive sentiment predictions with confidence scores.
• Interactive UI: Built with Streamlit for an intuitive and responsive user experience.
• Visualizations: Incorporates Plotly for creating interactive charts to display model performance and prediction confidence.

Prerequisites

• Python 3.7+
• pip (Python package installer)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/sentiment-analysis-streamlit.git
   cd sentiment-analysis-streamlit
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

Usage

To run the Streamlit app:

streamlit run app.py

This will start a local server and open the app in your default web browser.

Navigation

The app consists of four main sections:

1. Home: Provides an introduction to the sentiment analysis application.
2. Train Model: Allows users to initiate the model training process.
3. Evaluate Model: Displays the performance metrics of the trained model using interactive bar charts.
4. Predict: Enables users to input text and receive sentiment predictions with confidence scores visualized using a gauge chart.

Project Structure

project_root/
├── README.md
├── requirements.txt
├── app.py
└── sentiment_analyzer/
    ├── __init__.py
    ├── config.py
    ├── dataset.py
    ├── model.py
    ├── train.py
    ├── evaluate.py
    └── predict.py

• app.py: The main Streamlit application file.
• sentiment_analyzer/: A package containing the core functionality for sentiment analysis.
  • config.py: Configuration settings for the project.
  • dataset.py: Handles dataset loading and preprocessing.
  • model.py: Defines the sentiment analysis model.
  • train.py: Contains the training logic.
  • evaluate.py: Implements model evaluation.
  • predict.py: Provides functionality for making predictions.

Customization

You can customize various aspects of the project by modifying the config.py file:

• DATASET_NAME: Change the dataset used for training.
• NUM_SAMPLES: Adjust the number of samples used from the dataset.
• MODEL_NAME: Use a different pre-trained model.
• MAX_LENGTH: Modify the maximum sequence length for tokenization.
• BATCH_SIZE, LEARNING_RATE, NUM_EPOCHS: Tune the training hyperparameters.

Contributing

Contributions to this project are welcome! Please fork the repository and submit a pull request with your proposed changes.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Hugging Face for their transformers and datasets libraries.
• Streamlit for the amazing web app framework.
• The creators and contributors of the IMDB dataset.

Contact

For any questions or feedback, please open an issue in the GitHub repository or contact ME Here.