Error with finding onnxruntime_binding.node on Windows 10 on a bootcamp Macbook

[samlhuillier]

Describe the issue

When I try to run some basic onnxruntime code in a basic node 18 project on Windows 10 bootcamp on a macbook , I get the following error:

App threw an error during load
Error: The specified module could not be found.
\\?\C:\Users\samto\Desktop\reor\node_modules\onnxruntime-node\bin\napi-v3\win32\x64\onnxruntime_binding.node
    at process.func [as dlopen] (node:electron/js2c/asar_bundle:2:2238)
    at Module._extensions..node (node:internal/modules/cjs/loader:1356:18)
    at Object.func [as .node] (node:electron/js2c/asar_bundle:2:2238)
    at Module.load (node:internal/modules/cjs/loader:1126:32)
    at Module._load (node:internal/modules/cjs/loader:967:12)
    at l._load (node:electron/js2c/asar_bundle:2:13642)
    at Module.require (node:internal/modules/cjs/loader:1150:19)
    at require (node:internal/modules/cjs/helpers:119:18)
    at Object.<anonymous> (C:\Users\samto\Desktop\reor\node_modules\onnxruntime-node\dist\binding.js:9:1)
    at Module._compile (node:internal/modules/cjs/loader:1271:14)

To reproduce

Create a basic node 18 project on windows 10 bootcamp machine from a macbook. And try running the following code:

const ort = require("onnxruntime-node");

async function main() {
  try {
    // Load the ONNX model (make sure to provide the correct path to your model)
    const session = await ort.InferenceSession.create(
      "C:UserssamtoDownloadssqueezenet1.0-7.onnx"
    );

    // Create a tensor for input (adjust the size and data according to your model's requirements)
    // Example: Creating a dummy input tensor
    const inputTensor = new ort.Tensor(
      "float32",
      new Float32Array(1 * 3 * 224 * 224),
      [1, 3, 224, 224]
    );

    // Run the model with the tensor (replace 'inputName' with your model's actual input name)
    const feeds = { inputName: inputTensor };
    const results = await session.run(feeds);

    // Process the output (replace 'outputName' with your model's actual output name)
    const outputTensor = results.outputName;
    console.log(`Model output: ${outputTensor.data}`);
  } catch (error) {
    console.error("Error during ONNX session:", error);
  }
}

main();

Urgency

No response

Platform

Windows

OS Version

10

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.16.3

ONNX Runtime API

JavaScript

Architecture

X64

Execution Provider

Default CPU

Execution Provider Library Version

No response

[wschin]

npm install onnxruntime-node if your machine doesn't have it.

[samlhuillier]

Yes I do do that. The above error occurs after doing the npm install you mention...

[maani7823]

Hi @samlhuillier you need to install this https://aka.ms/vs/17/release/vc_redist.x64.exe

onnxruntime-node require you to have MSVS C++ redist

[A5an]

Hi @samlhuillier you need to install this https://aka.ms/vs/17/release/vc_redist.x64.exe

onnxruntime-node require you to have MSVS C++ redist

does that will work on the deployed server after that? Like, should I install or configure something on the vercel, to make sure that onnxruntime-node would work?