Play Rock-Paper-Scissors against a smart AI that learns your patterns!
This project uses Meta's LLaMA 3 model served locally through Ollama, wrapped in a clean and interactive Streamlit web app.
- Tracks your move history (rock/paper/scissors count).
- Sends a prompt to LLaMA 3 via Ollama to predict the best AI move.
- Decides a winner and updates the score dynamically.
- Runs entirely locally — no cloud or API keys needed.
| Tech | Description |
|---|---|
| 🐍 Python | Core game logic and integration |
| 🦙 Ollama | Local LLaMA 3 model server |
| 🧠 LLaMA 3 | Meta’s advanced language model |
| 🌐 Streamlit | Web interface for easy interaction |
git clone https://github.com/your-username/rock-paper-scissors-ai-ollama.git cd rock-paper-scissors-ai-ollama
Install dependencies:
pip install -r requirements.txt
Install and run the model locally:
curl -fsSL https://ollama.com/install.sh | sh
ollama pull llama3 ollama run llama3
Keep this terminal running while playing.
streamlit run rps_app.py
├── main.py # CLI version of the game
├── rps_app.py # Streamlit web app
├── game_logic.py # AI integration & logic
├── requirements.txt # Dependencies
├── .gitignore # Ignored files
├── LICENSE # MIT license
└── README.md # This file
- ✅ Fun and intelligent gameplay
- 🧠 Learns from your move history
- ⚡ Fast and runs offline
- 🛠️ Easy to customize and extend
- 🌍 No API keys, no external dependencies
“You are playing rock-paper-scissors. The player has played 'rock' 3 times, 'paper' 1 time, and 'scissors' 6 times. Based on this, what move would you choose next to win?”
This project is licensed under the MIT License. LLaMA 3 is provided and governed by Meta's licensing terms. Ollama usage is governed by Ollama's terms.
Sagnik Datta Built with ❤️ using AI tools. 🔗 GitHub