Tensorflow.js Node.js Examples. Inspired from official tfjs-examples.
Tensorflow.js for Node.js (namely tfjs-node) is currently in development status, see here.
Because officiale Tensorflow.js examples are mostly written in TypeScript or use the async/await ES2017 paradigma, while the most of node modules are not.
Tf.js Hello World for Node.js: training and predict.
Load and predict sentiment from a pre-trained sentiment model with tf.loadModel.
A simple LSTM train example that shows accuracy, loss, while training.
A bilinear LSTM (bi-LSTM) for EN-FR translation with a local hosted saved model. Full examples: load model, translate sentences.
A bi-LSTM model that learns to write as Nietzsche or your favorite text. Work in progress, so far it loads the saved model only.
An example that show how to load a Tensorflow SavedModel, without any conversion, using tf.node.loadSavedModel. It loads the model, predict the bounding boxes of detection objects and write to an image using a canvas.
Yes! Definitively you can. See async2promise.js. You have here defined two function the __awaiter and the __generator, that mimic the async and await using a Promise. Enjoy ECMA5.