Inference

Launching demo application

As the training of the model from previous step continues, launch the Custom Labels demo application. We will use this web application to test the custom model.

You can use click on “Launch on AWS” button to launch the stack in your account. Make sure you use appropriate region.

Amazon Rekognition Custom Labels Demo

Wait for the training to complete before you start with the following steps.

Inference

Navigate to the project you trained your model under, wait for the status to show TRAINING_COMPLETE and then click the on the model Name of the trained model:

Evaluate your model performance details.

If you’re happy with the performance of your model, you make it available for use by starting it.

Inference using GUI-based application

In this approach, we will use the demo application we launched at the beginning of the section. Follow the steps below:

  1. Open AWS console and switch to “us-east-1” region.
  2. Under services search for “Amazon CloudFormation”.
  3. Find the stack you launched for the demo application.
  4. Go to the “Outputs” tab and copy the URL.
  5. Paste the URL in a browser window.
  6. You would have received the email with credential for the demo application. Use them to log in. You will be asked to change the password.

Once logged in, you will be presented with a page similar to below.

Click on “Start the model” for the model you just trained. Wait for the model to reach “RUNNING” state. Click on the model name. You will see a page similar to below.

Use the following images to test your model.

Upload the images one after the other to see if it can identify AWS logo.

If the logo is detected, it will draw a bounding box around the logo. It will also show the confidence score under “Results”.

Click on “Projects” to go back to main page. “Stop the model” once you’re done with testing.

Inference using AWS CLI

Scroll down to the Use your model section. Expand the section and copy code.

You can use “Terminal” on the SageMaker Notebook you launched in Lab 1 to run the commands.

Alternatively, you can use AWS CLI on your local machine. To install AWS CLI, follow - https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-install.html

After you’ve started your model, you can use it to analyze images.

Use the following images to test your model.

You are charged for the amount of time your model is running. If you have finished using the model, you should stop it.