As mentioned earlier, this can be important information for data scientists to know before they export a project, as they can use it to understand data distribution. In this case, if we wanted to train a model that can detect all types of candies including "Dumle", we would need to add more annotations for that class before exporting.