For example, in default Tensorflow training configuration you can change batch size and number of iterations. For start creating Neural Network press run button and select training location.
You can train your model on local machine for free or in cloud with some benefits.
For local training you should install Docker. You can do it with https://hub.docker.com/editions/community/docker-ce-desktop-mac. Also you should keep MakeML active all time during training.
For training in cloud you should keep MakeML active only during archiving and uploading dataset.
After selecting training location you will see “Trainings” window, also you can find it by Window -> Trainings. There you can track status and get result of your training.
When your training will be finished “Export” button will be available. If your training was on local machine just press “Export model” button and choose directory.
For exporting model from training in cloud, press “Pay and export button” and wait until your model will be downloaded. Please, be patient, sometimes it can take a lot of time. Then you should enter your credit card info and press “Pay” button.
Then you should select directory for exporting your model.
Looks like you are ready for using model in your application. You can ready how to use CoreML model in iOS application in our documentation.