search

Solacex / CCM

[ECCV2020] Content-Consistent Matching for Domain Adaptive Semantic Segmentation
MIT License
94 stars 10 forks source link

thres in gene_thres #6

Closed Lufei-github closed 3 years ago

Lufei-github commented 3 years ago

A small question!

Your ccm_trainer.py

159 index = int(cls_total * prop) 160 cls_thres = cls_prob[-index] 161 cls_thres2 = cls_prob[index] 162 thres[k] = cls_thres

For what consideration, you set your thres be the value of the last ten percent of the probability value?

last bacause of (-index), ten percent because of (prop=0.1)

Solacex commented 3 years ago

Hello, to obtain the pseudo label, we adopt a thresholding strategy where only the top 10% samples of each class are assigned with pseudo labels.