search

XiaLiPKU / EMANet

The code for Expectation-Maximization Attention Networks for Semantic Segmentation (ICCV'2019 Oral)
https://xialipku.github.io/publication/expectation-maximization-attention-networks-for-semantic-segmentation/
GNU General Public License v3.0
680 stars 130 forks source link

Fail to reproduce your results on COCO-STUFF #23

Closed zenght closed 4 years ago

zenght commented 4 years ago

Hi @XiaLiPKU ! Your work is amazing, and i am appreciate that you have released your code. May i ask question? Based on your code, i have modified your code to suit COCO-STUFF training. But i can only get 34.55% miou. I just followed your default setting, but in single gpu. ( pretrained ResNet-101, batch size 3, 30k iterations and so on...).

Looking forward to your reply! Best wishes

XiaLiPKU commented 4 years ago

Hi @XiaLiPKU ! Your work is amazing, and i am appreciate that you have released your code. May i ask question? Based on your code, i have modified your code to suit COCO-STUFF training. But i can only get 34.55% miou. I just followed your default setting, but in single gpu. ( pretrained ResNet-101, batch size 3, 30k iterations and so on...).

Looking forward to your reply! Best wishes

Batch size is one of the most factor of segmentation's performance. In my experiment, the batch size is 16