Appropriate recipes Multilabel Classification

Hi, I want to do multilabel classification (non-exclusive multiclass) and as I found in PRODIGY_README.html I have to use choice recipe with exclusion flag turned off.
Can you somehow share steps needed for such task like (teach terms for each class, use some recipe, label data for all class together or separately, which recipe to use for multilabel classification, how multilabel classifier model works (does it have softmax and normalize class probabilities or not, etc).
Thanks in advance!

This may give you a head start: