Developing intelligent machine leaning models for solving a particular problem with considerable accuracy in itself is a great challenge. We can manage to build the most optimum model, but unless we know how to put it into production, it’s hard to get it to create the maximum amount of possible value. This is achieved by setting up the model in the production based environment through the use of certain explicit APIs and web application which forms an intermediary between the user and the model. There are platforms that serves the purpose of deploying machine learning models, some of them are – Google cloud service, Amazon web service (AWS) etc.

Handwritten Digit Classification model using Tensorflow JS

TensorFlow.js is one of such library used for deploying machine learning using JavaScript. We can explore the visualization demo for seeing the training process of the Digital handwritten classification model by clicking on the below link to visualize the training process of the Digital handwritten classification model.

Handwritten Digit Classification model deployment test

When the training finished, you can draw a digit on the black canvas and see if the model is predicting the handwritten digit accurately.

Note: This may not work on the mobile/tablet devices.

Useful Resources

TensorflowJS

Kubernetes Basics

Kubeflow

Continuous Delivery

Machine Learning Model Management