FractalAI / README.md
Aluode's picture
Update README.md
b2b7236 verified
---
title: FractalAI
emoji: πŸ†
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.1.0
app_file: app.py
pinned: false
license: mit
short_description: Crazy experiment
---
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
# Fractal AI with Visualization and Interaction
## Description
This project implements an advanced Fractal AI system with dynamic growth, visualization, and interactive features. It combines concepts from fractal geometry, neural networks, and Hebbian learning to create a unique and evolving AI structure.
**Project Video Demonstration:**
[![Fractal AI Project Video](https://img.youtube.com/vi/M1mV1a5Te44/0.jpg)](https://www.youtube.com/watch?v=M1mV1a5Te44)
## Features
- Dynamic fractal network growth
- 3D visualization of the fractal AI structure
- Hebbian learning for connection weight updates
- Interactive chat functionality
- Wikipedia integration for training
- Self-conversation mode
- State saving and loading
- Zoom functionality for detailed exploration
## Installation
1. Clone this repository:
```
git clone https://github.com/yourusername/fractal-ai.git
cd fractal-ai
```
2. Install the required dependencies:
```
pip install -r requirements.txt
```
## Usage
Run the main script:
```
python fractal_ai.py
```
This will launch a Gradio interface in your default web browser, where you can interact with the Fractal AI system.
## Interface Options
- **Run Cycles**: Execute a specified number of thinking cycles
- **Train on Wikipedia**: Input a topic to train the AI on Wikipedia content
- **Chat**: Engage in a conversation with the AI
- **Self-Conversation**: Let the AI converse with itself
- **Zoom**: Explore the fractal structure in detail
- **Save/Load State**: Preserve or restore the AI's state
## Contributors
- Antti Luode - Original concept and ideation
- ChatGPT - Assisted in code generation and problem-solving
- Claude (Anthropic) - Implemented core functionality and resolved issues
## Acknowledgements
Special thanks to Antti Luode for the innovative and ambitious idea behind this project. The collaboration between human creativity and AI assistance has made this unique project possible.
## License
This project is open-source and available under the MIT License.