Mawiszus/TOAD-GUI
🔧 ToolMawiszus
A user-friendly GUI for TOAD-GAN, enabling procedural generation of token-based Super Mario Bros. levels.
TOAD-GUI serves as the front-end implementation for the TOAD-GAN model, which utilizes a token-based approach to learn and generate level structures from Super Mario Bros. The tool leverages deep learning techniques to understand the spatial and logical constraints of platformer levels. Key features include an intuitive interface for configuring generation parameters, real-time visualization of generated level segments, and support for token-based level representation. By abstracting the underlying GAN training and inference processes, the GUI allows users to experiment with different latent spaces and generation seeds without needing to write extensive boilerplate code. It is a vital tool for those exploring the intersection of generative adversarial networks and game design, specifically focusing on how neural models can replicate the design patterns found in classic platforming games.
💡Highlights
- ├─GUI for TOAD-GAN architecture
- ├─Token-based level generation
- └─PCG-ML for platformer games
🎯For
- ├─Game developers
- └─AI researchers