This example shows how to predict the age of abalone from physical measurements using TensorFlow.js with Node.js.

The data set available at UCI Machine Learning Repository.

This example shows how to:

• load a Dataset from a local csv file.
• prepare the Dataset for training.
• create a tf.LayersModel from scratch.
• train the model through model.fitDataset().
• save the trained model to a local folder.

To launch the demo, run the following command:

yarn
yarn train

The result logs 100 Epochs as well as a predicted result similar to the following:

...
Epoch 100 / 100
eta=0.0 =================================================>
402ms 57414us/step - loss=7.42 val_loss=5.60
The actual test abalone age is 10, the inference result from the model is 11.929240226745605

By default, the training uses tfjs-node, which runs on the CPU. If you have a CUDA-enabled GPU and have the CUDA and CuDNN libraries set up properly on your system, you can run the training on the GPU by replacing the tfjs-node package with tfjs-node-gpu.