search

LiheYoung / UniMatch

[CVPR 2023] Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
https://arxiv.org/abs/2208.09910
MIT License
478 stars 60 forks source link

a question about data split #1

Closed Jack-bo1220 closed 2 years ago

Jack-bo1220 commented 2 years ago

A nice work! Is the data split of the Cityscapes dataset used the same as that of the previous work (i,e,. ST++)? I noticed that there were many splits in the previous work. Is it the same as theirs? Thank you!

LiheYoung commented 2 years ago

Thanks for your interest.

We have four data partitions on the Cityscapes. The 1/16, 1/8, and 1/4 partitions are the same as CPS, while the 1/30 partition is the same as ST++, because CPS did not conduct the 1/30 setting.