README

Microsoft.ML.OnnxRuntime.EP.WebGpu

WebGPU plugin Execution Provider for ONNX Runtime.

Prerequisites

This package provides the WebGPU plugin EP only. Your project must separately reference an ONNX Runtime core package (e.g. Microsoft.ML.OnnxRuntime) of version @min_onnxruntime_version@ or later.

If the referenced ONNX Runtime is incompatible, the plugin EP will report an error when its library is registered.

On Linux, a system Vulkan loader (libvulkan.so.1) must be installed and available at runtime.

Supported Platforms

Installation

dotnet add package Microsoft.ML.OnnxRuntime --version @min_onnxruntime_version@
dotnet add package Microsoft.ML.OnnxRuntime.EP.WebGpu

Usage

// Note: Error handling is omitted for brevity, except for the device-discovery check below.

using Microsoft.ML.OnnxRuntime;
using Microsoft.ML.OnnxRuntime.EP.WebGpu;

// Register the WebGPU EP plugin library
var env = OrtEnv.Instance();
env.RegisterExecutionProviderLibrary("webgpu_ep", WebGpuEp.GetLibraryPath());

// Find the WebGPU EP device
OrtEpDevice? webGpuDevice = null;
foreach (var d in env.GetEpDevices())
{
    if (d.EpName == WebGpuEp.GetEpName())
    {
        webGpuDevice = d;
        break;
    }
}
if (webGpuDevice is null)
{
    throw new InvalidOperationException("No WebGPU device found.");
}

// Create a session with the WebGPU EP
using var sessionOptions = new SessionOptions();
sessionOptions.AppendExecutionProvider(env, new[] { webGpuDevice }, new Dictionary<string, string>());

using var session = new InferenceSession("model.onnx", sessionOptions);

Troubleshooting

• No WebGPU device found — the plugin EP loaded but no compatible adapter was discovered. On Linux this usually means the Vulkan loader (libvulkan.so.1) is not installed; install it via your distribution's package manager. On Windows it may indicate a missing or outdated GPU driver.
• ORT runtime version "..." is below the minimum required version "@min_onnxruntime_version@" — the referenced Microsoft.ML.OnnxRuntime package is older than @min_onnxruntime_version@. Upgrade with dotnet add package Microsoft.ML.OnnxRuntime --version @min_onnxruntime_version@.