nodejs/onnxruntime.js

const fs = require('fs');
const sharp = require('sharp');
const ort = require('onnxruntime-node');

(async () => {
  try {
    // Step 1: Load and preprocess the image
    const imageBuffer = await sharp('./training_images/hat.jpg')
      .resize(128, 128) // Resize to 128x128
      .raw() // Get raw pixel data
      .toBuffer();

    // Normalize to [0, 1] and create input tensor
    const imgArray = Float32Array.from(imageBuffer).map(value => value / 255.0);
    const inputTensor = new ort.Tensor('float32', imgArray, [1, 128, 128, 3]);

    // Step 2: Load ONNX model
    const session = await ort.InferenceSession.create('./saved-model/model.onnx');

    // Step 3: Run inference
    const results = await session.run({ [session.inputNames[0]]: inputTensor });

    // Step 4: Get output probabilities
    const probabilities = results[session.outputNames[0]].data; // Float32Array
    const labelIndex = probabilities.indexOf(Math.max(...probabilities)); // Find the index of the max probability

    // Step 5: Load the label map
    const labelMap = JSON.parse(fs.readFileSync('./labelMap.json', 'utf8')); // Assuming you saved the label map
    const label = Object.keys(labelMap).find(key => labelMap[key] === labelIndex);

    console.log(`Predicted label: ${label}`);
    console.log(`Confidencex: ${(probabilities[labelIndex] * 100).toFixed(2)}%`);
  } catch (err) {
    console.error('Error:', err);
  }
})();