🤖 Q&A Demo (Streamlit + Hugging Face)

A minimal Q&A web app built with Streamlit that calls the Hugging Face Inference API to generate answers from open-source LLMs.
It first tries a chat model and automatically falls back to a text-generation model if needed.

---

✨ Features

• 🖥 Streamlit UI – Clean and interactive interface
• 🔑 Secure API Authentication – Uses Hugging Face Access Tokens via environment variables
• 🧠 Two-stage model flow:
  1. HuggingFaceH4/zephyr-7b-beta via chat_completion
  2. Fallback: tiiuae/falcon-7b-instruct via text_generation
• 🚦 Error Handling – Displays clear error messages and status info
• 🌐 Deployable to Hugging Face Spaces with minimal setup

---

📂 Project Structure

.
├── app.py               # Main Streamlit application code
├── requirements.txt     # Python dependencies needed to run the app
├── README.md            # Project documentation

---

🚀 Deploy to Hugging Face Spaces

1. Create a Space: https://huggingface.co/spaces
   - SDK: Streamlit
   - Runtime: Python 3.10+

2. Push code to Space repository
   git remote add space https://huggingface.co/spaces/<username>/<space-name>
   git push space main

3. Add your API token as a secret
   - Go to Space → Settings → Repository secrets
   - Add HUGGINGFACEHUB_API_TOKEN with your Hugging Face token value

4. Space will build and deploy automatically.

---

🧠 How It Works

1. User enters a question in the input box.
2. App verifies Hugging Face token via HfApi().whoami().
3. Attempts to answer with chat_completion using Zephyr-7B.
4. If chat model fails, falls back to text_generation using Falcon-7B.
5. Displays the answer in the UI.

---

🛠 Configuration

Environment Variable:
- HUGGINGFACEHUB_API_TOKEN (required, must start with hf_)

Models:
- Primary: HuggingFaceH4/zephyr-7b-beta
- Fallback: tiiuae/falcon-7b-instruct

Parameters:
- max_tokens / max_new_tokens: control output length
- temperature, top_p: adjust randomness

---

🩹 Troubleshooting

Token Error:
- Ensure your token is valid and starts with hf_
- Check that it's exported in the same shell session

Model Error:
- Switch to another available model on Hugging Face
- Check model’s Inference API page to confirm it supports chat or text generation

---

🤝 Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/YourFeature)
3. Commit your changes (git commit -m "Add some feature")
4. Push to the branch (git push origin feature/YourFeature)
5. Open a Pull Request

---