gonet

GoNet: Lightweight Neural Network in Go

GoNet is a lightweight, minimalistic neural network implementation written in Go. Designed with simplicity in mind, it currently supports multilayer perceptrons (MLPs) and uses batch stochastic gradient descent with backpropagation for training.

Features

• ✔️ Core Functionality: Implements basic MLP architecture.
• ✔️ Training Algorithm: Batch stochastic gradient descent with backpropagation.
• ✔️ Lightweight Design: Focused on simplicity and clarity, making it an excellent starting point for those interested in learning about neural networks.

Current Limitations

GoNet is still under development and lacks the following features:

• ❌ CSV Data Reading: Built-in support for loading data from CSV files.
• ❌ Model Execution Post-Training: The ability to load a saved model and run it for inference.

These features may be added in the future, but active development may be limited due to other commitments.

Usage

To get started with GoNet, you can install the library using the following command:

go get github.com/AdamJacuch/GoNet@latest

Sample Implementation

Below is an example of how to use GoNet to train and test a simple neural network:

package main

import (
	"fmt"

	gonet "github.com/AdamJacuch/GoNet"
)

func main() {
	gonet.NumEpochs = 1000
	gonet.BatchSize = 4

	data := [][]float64{
		{0, 0, 1, 1, -1, 1, 0, -1},
		{1, 1, 1, 1, -1, 1, 0, 1},
		{-1, 0, -1, 1, -1, -1, -1},
		{-1, 1, -1, 1, -1, 1, -1},
		{1, -1, 1, -1, 0, 0, 0, 0},
		{1, 1, 0, 0, 0, 0, 1, 1},
		{-1, -1, 0, 0, 0, 0, -1, -1},
		{0, 0, 0, 0, 0, 0, 0, 0},
		{1, 1, 1, 1, 1, 1, 1, 1},
		{0, 0, 0, 0, 0, 0, 0, 0},
		{-1, -1, -1, -1, -1, -1, -1, -1},
		{1, 1, 0, 0, 0, 0, -1, -1},
	}
	outputs := [][]float64{
		{0, 1, 0},
		{0, 0, 1},
		{1, 0, 0},
		{0, 1, 0},
		{0, 1, 0},
		{0, 0, 1},
		{1, 0, 0},
		{0, 1, 0},
		{0, 0, 1},
		{0, 1, 0},
		{1, 0, 0},
		{0, 1, 0},
	}

	nn := gonet.NewNet([]int{8, 5, 3})

	nn.Train(data, outputs, 0.01, 10)

	output := nn.ForwardPass([]float64{-1, -1, 1, 1, 1, 1, 1, 1}) // should be {0, 0, 1}
	gonet.Softmax(&output)
	fmt.Println(output)
}

Contributions

Contributions are welcome!

• Feel free to submit a pull request.
• Open an issue to suggest new features or report bugs.

Disclaimer

This project is a work in progress and may not be actively maintained.
Use it as-is or as a foundation for your own projects.