🤖 NanoNeuron

NanoNeuron is seven simple JavaScript functions that give you a feel for how a machine can learn. There are no libraries, no external datasets and no dependencies, just plain functions for prediction, cost, forward and backward propagation, and training.

The NanoNeuron learns a single task: converting temperatures from Celsius to Fahrenheit. It starts with random parameters and adjusts them step by step until its predictions match the real formula, which shows that machine learning is math, not magic.

Many concepts are deliberately simplified or skipped, so it's a first step toward understanding neural networks rather than a complete guide.