|
|
--- |
|
|
license: apache-2.0 |
|
|
title: 'GUVI AI ' |
|
|
emoji: 🌍 |
|
|
colorFrom: green |
|
|
colorTo: blue |
|
|
sdk: streamlit |
|
|
sdk_version: 1.49.1 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
short_description: Learning Assistant |
|
|
--- |
|
|
|
|
|
# GUVI Multilingual AI Chatbot |
|
|
|
|
|
This project is a sophisticated multilingual chatbot built with Streamlit and powered by Hugging Face Transformers. It's designed to understand and respond to user queries in over 20 languages, providing real-time, domain-specific answers about GUVI's courses. |
|
|
|
|
|
## 🚀 About This Project |
|
|
|
|
|
The primary goal of this chatbot is to assist learners on platforms like GUVI by providing real-time, multilingual support. It helps break down language barriers for non-English-speaking users, making educational resources more accessible. |
|
|
|
|
|
## ⚙️ How It Works |
|
|
|
|
|
The chatbot now features a fully automated and intelligent workflow: |
|
|
|
|
|
1. **Auto-Detect Language**: When a user enters a query, the app first uses a `language-detection` model to instantly identify the input language (e.g., Tamil, Hindi, Spanish). |
|
|
2. **Translate to English**: If the detected language is not English, the query is translated into English using the `facebook/mbart-large-50-many-to-many-mmt` model. |
|
|
3. **Generate Accurate Response**: The English query is passed to a custom **fine-tuned GPT-2 model** (`aswinprasath31/guvi-gpt2-finetuned`). This model has been specifically trained on GUVI's data to provide relevant and accurate answers about courses, enrollment, and more. |
|
|
4. **Translate Back**: Finally, the English response is translated back into the user's originally detected language before being displayed in a clean, chat-like interface. |
|
|
|
|
|
## 🛠️ Tech Stack |
|
|
- Python |
|
|
- Streamlit (for the web interface) |
|
|
- Hugging Face Transformers (for AI models and fine-tuning) |
|
|
- PyTorch |
|
|
- Google Colab (for GPU-accelerated model training) |
|
|
|
|
|
## 📋 How to Run This Project |
|
|
|
|
|
1. **Clone the Repository** |
|
|
```bash |
|
|
git clone [https://github.com/Aswinprasath31/GUVI-Multilingual-GPT-Chatbot.git](https://github.com/Aswinprasath31/GUVI-Multilingual-GPT-Chatbot.git) |
|
|
cd GUVI-Multilingual-GPT-Chatbot |
|
|
``` |
|
|
|
|
|
2. **Install Dependencies** |
|
|
```bash |
|
|
pip install -r requirements.txt |
|
|
``` |
|
|
|
|
|
3. **Run the Streamlit App** |
|
|
```bash |
|
|
streamlit run app.py |
|
|
``` |
|
|
The application will then be running and accessible in your web browser. |
|
|
|
|
|
## 🔗 Deployed Application |
|
|
|
|
|
You can interact with the live chatbot deployed on Hugging Face Spaces here: |
|
|
|
|
|
[**https://huggingface.co/spaces/aswinprasath31/guvi-multilingual-chatbot**](https://huggingface.co/spaces/aswinprasath31/guvi-multilingual-chatbot) |
