problem of Subclass split

[xiewende]

nice works！
i want to know the problem of sub-class split
for example：self.split_list = [132, 2, 1, 1, 1, 2, 3, 4, 7, 2, 1, 2, 6, 2, 2, 15, 1, 1, 2, 2, 1] in clustering.py，it means that original-class 0 split into 132 sub-class，however，how to get this 132 sub-class？ Is the sub-class divided according to the size of the target area of ​​the original-class?

[Dayan-Guan]

Hi @xiewende , thank you for your comments.
The subclass split is set according to the ratio of each original class in the labeled data.

[xiewende]

for example，in one image ，caontains original-class 0，Is this class divided into 132 subclasses by k-means？If so，Wouldn't the target region of each subclass be tiny？

[Dayan-Guan]

Hi @xiewende , we perform clustering on the whole dataset.