Label vs. Class

[vapaunic]

There is no consistency on how terms label and class are used. I expect what you want is this hierarchy: single label can have many classes. However, occasionally you use label to imply class. For example in the following two sentences in 01_training_introduction notebook:

• Each folder title corresponds to the image label for the images contained inside:
• The widget shows test images along with the ground truth label and model prediction score.

Consider double checking all notebooks and other material to make sure you use these terms consistently.

[loomlike]

@vapaunic @jiata @PatrickBue Good point. Not sure if I understand this statement though "single label can have many classes". Correct me if I'm thinking wrong:

• When we say 'labeling', we refer an action of labeling each object. We can label one class or multiple classes to a single object, depending on the problem.
• When we call 'class label', it is the class name.
• When we call 'multi-label problem', it is the problem which a single object have more than two class-labels.
• When we call 'multi-class problem', the classification problem consists of multiple classes.

FYI, in scikit, https://scikit-learn.org/stable/modules/multiclass.html

Multiclass classification makes the assumption that each sample is assigned to one and only one label: a fruit can be either an apple or a pear but not both at the same time. Multilabel classification assigns to each sample a set of target labels