Training an object detection model using Azure AutoML
In this section, we will show you how to use Azure Automated Machine Learning (AutoML) to train an object detection model by following these steps:
- Click on Automated ML from the left navigation bar and then click New Automated ML job.
Figure 10.18 – Creating a new Automated ML job
- Select the image data asset that you created in the last section and click Next.
- In this step, you will configure the training job, as shown in Figure 10.19. The first thing to notice is that AutoML has automatically set the task type to Object Detection based on the selected dataset. Since you don’t have an existing experiment, select Create new to create a new experiment and call it
playing_cards_experiment. For Target column, select label (List), and for the compute type, select Compute cluster. Since you do not have a GPU cluster (which is needed for deep learning tasks such as object...
