Cascade Lattice Demo - Sovereign LunarLander
A standalone Windows executable demonstrating cascade-lattice provenance tracking with a trained RL agent playing LunarLander.
๐ Quick Start
- Download
cascade-demo-windows-x64.zip - Extract anywhere
- Run
CascadeDemo.exe
No Python or dependencies required!
๐ฎ What It Does
Watch a PPO-trained agent land a lunar module while cascade-lattice tracks:
- Model provenance - cryptographic identity of the trained model
- Inference observations - every decision the agent makes
- Reward attribution - linking outcomes to model behavior
๐ง Controls
- SPACE - Toggle pause
- R - Reset episode
- ESC - Quit
๐ฆ What's Inside
- Pre-trained PPO agent (downloaded from HuggingFace on first run)
- Pygame visualization
- Cascade provenance logging (optional)
๐ Links
- Source Code: tostido/Cascade on HuggingFace Spaces
- PyPI Package:
pip install cascade-lattice - Build Scripts: See
build/folder in the source repo
๐ System Requirements
- Windows 10/11 (x64)
- ~300MB disk space
- No GPU required
๐๏ธ Build It Yourself
cd Cascade_Repo/build
pip install -r requirements-build.txt
python build_exe.py
Built with cascade-lattice - Provenance tracking for AI systems
- Downloads last month
- 7