Skip to main content Link Menu Expand (external link) Document Search Copy Copied
allyoushawn's homepage

What’s this

Table of contents

What’s this about?

It’s a project building a machine learning service from scratch. From the model training to its deployment. From building backend REST API to frontend interface. It’s plug and play. You could just replace the model while leave the ml service untouched to run different ML services.

This project is suitable for people who want to have a quick ML service POC for the models they worked on. Enjoy :wine_glass:

A quick demo

Hit the endpoint allyoushawn.hopto.org:8080/ml_service through a browser, we could start using the ML service.

ml_service_home_page

In the page, we see the following elements:

  • Id: The task id specified by the user.
  • Message: The input of the ML service. Generally it’s a sentence.
  • A service selection menu: Currently we could choose sentimentAnalysis and machineTranslation.

After we click Submit, the service would give users the corresponding results. In this example, the sentiment score for the sentence This is very good. is 3.0. (which is ranging from -5.0 to 5.0. ) We will talk more about it in sentiment analysis section.

ml_service_sa_result

Let’s try the other service, machineTranslation. We could see the translated results in Chinese after we submit our query to the service.

ml_service_click_menu

ml_service_mt_result

However, just a disclaimer, you would find weird results when you input different sentences since the service is only a POC for the serving pipeline. We could spend tons of time tuning the ML models if needed. :stuck_out_tongue:

Repositories included in the projects

Here is the list of repositories: