ML5-Blazor

An Easy Machine Learning Library for Blazor.

Now supports both Blazor Server and WASM and MAUI Hybrid.

Current Features

1. Neural Network
2. Image Classification
3. Sound Classifier
4. Object Detector (YOLO and COCOSSD based)
5. PoseNet
6. Sentiment Analyzer
7. FaceMesh

Hosted Sample

Live Demo

Documentation

Install Asp.Net Core payload and then follow Installation Instructions here to configure ML5 to use it from C# Blazor app.
API documentation can be followed from here after lib configuration. .

Sample Neural Network

@page "/nn"
@using BlazorML5
@using BlazorML5.Helpers
@inject IJSRuntime runtime

<PageTitle>Index</PageTitle>
<button @onclick="AddData">Add Data and Train</button>
@code
{
    NeuralNetwork _network;
    protected override async Task OnInitializedAsync()
    {
        await Ml5.InitAsync(runtime);

        _network = await Ml5.NeuralNetworkAsync(new NeuralNetworkOptions()
        {
            Task = TaskType.Classification,
            DataUrl = "https://raw.githubusercontent.com/ml5js/ml5-library/main/examples/p5js/NeuralNetwork/NeuralNetwork_color_classifier/data/colorData_small.json",
            Debug = true,
            Inputs = new object[]{"r", "g", "b"},
            Outputs = new object[]{"label"}
        });
        _network.OnTraining+=(l,e)=>
        {
            Console.WriteLine($"Training: {e}%");
        };
        _network.OnTrainingComplete+=async ()=>
        {
            Console.WriteLine($"Training Complete");
            await _network.ClassifyMultipleAsync(new object[]{new object[]{12,13,14},new object[]{15,16,17}});
        };
        _network.OnDataLoaded+=(e)=>
        {
            Console.WriteLine($"Data Loaded");
        };
        _network.OnClassify+=async (l,e)=>
        {
            Console.WriteLine(e.Length);
            Console.WriteLine(e[0].Label);
        };
        _network.OnClassifyMultiple+=async (l,e)=>
        {
            Console.WriteLine(e.Length);
            Console.WriteLine(e[0].Length);
            Console.WriteLine(e[0][0].Label);
        };

    }
    async void AddData()
    {
        //data fetched from .csv file no need to manually add and hence directly training 
        //await _network.NormalizeDataAsync();
        await _network.TrainAsync();
    }

}

More samples here