This is a demonstration of running a machine learning model on a Blazor Web App. Here, the task is simply to classify images.
For image inference, the ONNX files for the SqueezeNet model trained on the ImageNet dataset are used.
Package and Version
- MudBlazor 6.0.17
- Microsoft.ML 1.7.1
- Microsoft.ML.ImageAnalytics 1.7.1
- Microsoft.ML.OnnxRuntime 1.13.1
- Microsoft.ML.OnnxTransformer 1.7.1
See BlazorMLWebApp.csproj for more details.
git clone https://github.com/napnel/BlazorMLWebApp.git
cd BlazorMLWebApp
dotnet watch run (or dotnet run)
- The development tool used is Visual Studio Code instead of Visual Studio.
- ML.NET does not seem to support blazor-wasm. Details are discussed in this isssue.
(It is said that the problem can be avoided by splitting the project into several projects)I will load the onnx through Javascript and infer.