Using the API is very easy, if you have a model that learns, and calculates x,y,z positons that you can send wia HTTP request (you can learn how to make a request here).You run the sample either in debug or release mode. 2 windows should pop up: one is a command line in the back, the other one is wehere everything will be displayed
After that, you need to post a request with the number of classes, and the file names paths.The needs the be in json format (here you can see an example from python). If your data went through you should see something similar to this.
In this example we trained a model on cat and dog images. So the x and y axes are named appropriately. Right now all the images are in the same place because we haven't sent any data about their positions yet. The data with the positions must contain the x, y, z values, in the same order as their file names.
If you everything went alright you should see similar thing happening on your screen. The model in this example is 85% accurate so you can crealry see that most of the images on the left are dogs, and on the fight are cats. The further away an image to the left or to the right the higher the confidence of the prediction is. In this example Hight corresponds to loss (it means how bad the prediction is), because of this the cat images on the left are higher.