Skip to content

Latest commit

 

History

History
122 lines (76 loc) · 3.45 KB

README.md

File metadata and controls

122 lines (76 loc) · 3.45 KB

AskFAST Chat Bot

Overview

🎓 AskFAST is a chat bot designed to handle admission-related queries for FAST. It utilizes the powerful Mistral-7B language model to provide accurate and intelligent responses.

Contributors

Repository Link

Requirements

  • Python 3.8 or later
  • CUDA enabled GPU (for local runs)
  • Google Colab (for cloud runs)
  • Gradio (for creating interactive web interfaces)
  • Vercel (for deployment)

Setup and Installation

  1. Clone the Repository

    git clone https://github.com/ahmedembeddedxx/AskFAST.git
cd AskFAST

  2. Install Dependencies

    pip install -r requirements.txt

  3. Ensure CUDA is enabled

    If running locally, make sure CUDA is properly installed and configured. For running in Google Colab, ensure the notebook is set to use a GPU runtime (preferably T4 GPU).

Running the Chat Bot

  1. Execute the API Script

    Navigate to the src/scripts/ directory and run the 5_AskFAST_API.py script:

    cd src/scripts
python 5_AskFAST_API.py

    Upon successful execution, a link similar to <random-string>.gradio.live will be generated.

  2. Update the Web Application

    Copy the generated link and paste it in the src/web-app/index.html file, replacing the placeholder in the <Button> send tag.

    <Button send="https://<random-string>.gradio.live">

  3. Run the Web Application

    You can run the application on localhost or deploy it on Vercel.

    • Localhost:

      Navigate to the web app directory and start a local server:

      cd src/web-app
python -m http.server

    • Vercel:

      Follow the Vercel deployment guide to deploy your application.

Data

A large amount of data is available in the src/data directory. This data was scraped using PyTesseract and is publicly available under the GNU and MIT licenses.

  • Data Directory: src/data
  • Scripts for Data Scraping: src/scripts/

License

The data used in this project is publicly available under the GNU and MIT licenses.

Acknowledgments

A big thanks to the following for their invaluable tools and support:

  • Unsloth for their data scraping services.
  • PyTesseract for optical character recognition.
  • Gradio for creating interactive web interfaces.
  • Hugging Face for providing the Mistral-7B language model.

Demonstration Video

Watch the demonstration video for a quick overview of the project:

Watch the video

Feel free to reach out to any of the contributors for questions or collaboration opportunities.