SvetLuna — NN Training Animation

Neural Network Training Visualization — an educational Python/Matplotlib demo illustrating decreasing loss, increasing accuracy, and a schematic network representation.

Author: Svetlana Romanova — SvetLuna Studio
LinkedIn: https://www.linkedin.com/in/svetlana-romanova-418596387

---

🌟 Overview

This project provides a concise visual demonstration of neural network training dynamics. It shows:

• a loss curve that decreases over epochs,
• an accuracy curve that increases,
• a simple network schematic whose connections “light up” as training progresses.

The goal is to give beginners and educators an intuitive, visual explanation of how models learn.

---

🔧 Python demo code

The repository contains a small pure-Python / Matplotlib demo:

• src/nn_training_animation.py — generates synthetic loss/accuracy curves and an animated schematic neural network;
• nn_training_animation.mp4 — an exported animation (example output);
• SvetLuna-NN-Animation-Project-ready.zip — original project bundle.

---

🚀 Quickstart

Create and (optionally) activate a virtual environment, then install dependencies:

pip install -r requirements.txt
Run the animation interactively

python -m src.nn_training_animation
Or render and save an MP4 (requires ffmpeg to be installed):

python -m src.nn_training_animation save
If saving fails (no ffmpeg), you can still watch the animation in a window.

📄 License & use
This repository is intended for education and self-study.
If you use the animation in your teaching materials or talks, a short mention
of the author (Svetlana Romanova — SvetLuna Studio) is welcome.