Some issues about the performance(mIoU)

YudeWang / SEAM

Self-supervised Equivariant Attention Mechanism for Weakly Supervised Semantic Segmentation, CVPR 2020 (Oral)

MIT License

539 stars 97 forks source link

Some issues about the performance(mIoU) #17

Closed ZhiyuXu0124 closed 3 years ago

ZhiyuXu0124 commented 3 years ago

Hi,

Thanks for sharing the code.

I am trying to reproduce your code, but the final result I got is 3% different from the result in your article.

I followed the steps in the readme exactly. The results of local training are Train: 63.420%, Val: 60.336%; the results obtained by using the model you provided are: Train: 63.606%, Val: 60.076%.

I am not sure where the problem occurred, and I look forward to your answer, thank you.

YudeWang commented 3 years ago

Hi @ZhiyuXu0124 ,

The results are ok, the mIoU of pseudo label is evaluated on train set which is mentioned in section4.3 (63.61%)

The results in Tab.7 is retrained segmentation model prediction on val set. Details can be viewed here: https://github.com/YudeWang/SEAM/issues/11

ZhiyuXu0124 commented 3 years ago

Ok, I see, thanks for your answer