AWS offers a number of products and solutions to support machine learning (ML). First, let's look at how to use AWS to launch a simple machine, following these steps:
- After logging into AWS, you will see the following screen:
- Select the Launch a virtual machine option. Next, you will see the following screen:
Getting started AWS screen
- Click on Get started to go to the next screen, as follows:
Entering your instance identity name
- Enter a name (probably your own name) to go to the next screen:
Selecting the operating system
- Select the operating system for deployment. Ubuntu is a good choice:
Selecting Ubuntu for deployment
- Select the type of instance, based on the hardware requirements:
- In order to log in to the machine, a key pair has to be created for security. Provide a name for the file:
Verifying the machine name
- You should see a status of Completed, as the preceding screenshot shows. Now, you can switch to the EC2 console by clicking on the following button:
- There will be a short pause, to launch the instance. Once the instance is created, an IP will appear for logging:
- By clicking on the Actions button and choosing the Terminate button, one can terminate the instance, whenever required:
By following the preceding steps, one can create an instance with the required hardware and deploy TensorFlow Serving for inference to be queried by the client.