The Cloud Annotations SDK makes it easy to use your custom trained object detection or classification models in the browser or on the backend with TensorFlow.js. Simply load() your model_web folder and pass the loaded model a <canvas>, <img> or <video> reference.
npm install @cloud-annotations/models
import models from '@cloud-annotations/models' const model = await models.load('/model_web')
const img = document.getElementById('img') const predictions = await model.detect(img) // predictions => [{ label: 'dog', bbox: [x, y, width, height], score: 0.92 }, { label: 'cat', bbox: [x, y, width, height], score: 0.72 }]
const img = document.getElementById('img') const predictions = await model.classify(img) // predictions => [ { label: 'dog', score: 0.92 }, { label: 'cat', score: 0.72 } ]
No npm install required. Just import via the script tag.
<script src="https://cdn.jsdelivr.net/npm/@cloud-annotations/models"></script> <script> const img = document.getElementById('img') models.load('/model_web') .then(model => model.detect(img)) .then(predictions => { console.log(predictions) }) </script>
Example usage: Real-Time Object Detection With React.