Take AutoML Vision for a test drive

What kind of cloud is this?

I don’t know either. Can you think of other types of images that would be useful to classify? For example, if you have an online tire shop, wouldn’t it be cool if your customers could upload a photo of their current tire to find a matching replacement?

AutoML can help your tire shop decrease the number of returns and increase customer delight. And it’s easier than you think – learn how to take advantage of this powerful new tool with a new lab, AutoML Vision API.

Here’s some of what you’ll do:

  • You’ll be prompted to log into the AutoML UI. Open the link in a new tab to make it easy to hop between AutoML and your lab manual. Then, you’ll log into AutoML with your Qwiklabs-provided credentials.
  • In the AutoML UI, you will see a button for “Billing”. You do not need to worry about billing for this lab, as Qwiklabs covers the cost of this lab for you.
  • When you load data into your storage bucket, you’ll set your project name with one of the first commands you run. Find your project name on the left of your lab manual, just below your GCP credentials:

AutoML is currently in beta, so we expect to see constant changes to the user interface. Send any questions to Support@Qwiklabs.com and we’ll be happy to help. Plus you’ll help us keep the lab up to date as the product continues to improve!

Curious about what type of clouds were spotted during our trip to Muir Woods? Here are our model’s predictions:

The model is 99% certain these are cumulus clouds! Prediction certainty and accuracy would be better if we trained our model with more than 20 images per category. When you’re helping customers order tires, you will probably want to use 100+ photos per category.

Happy lab-taking!