#Bipartition threshold used to train this model (only multi-label classification)
#Mon Sep 25 16:12:56 CEST 2023
threshold=0.5