AGCA icon indicating copy to clipboard operation
AGCA copied to clipboard

Published 20 hours ago • verified publisher icon C1nDeRainBo0M
→

Metadata

The core code of AGCA: An Adaptive Graph Channel Attention Module for Steel Surface Defect Detection

Results 4 AGCA issues
Sort by recently updated
recently updated
newest added

A0 needs to be placed on the GPU GPU和CPU

7 comment

作者您好，通过这步操作将A0放在GPU上， ![image](https://github.com/C1nDeRainBo0M/AGCA/assets/57579741/f0350ef5-6ed9-4d8b-b8cd-3b0f55a44af2) 但是其他模型和数据都还在cpu上 ![image](https://github.com/C1nDeRainBo0M/AGCA/assets/57579741/3c4b1508-a250-4b71-bcab-fa7996d11beb) 因此，在执行这步操作的时候： ![image](https://github.com/C1nDeRainBo0M/AGCA/assets/57579741/0b1270ba-b521-405e-9e8e-243ee73b8538) 就报错一个数据在GPU、一个数据在CPU，无法进行运算，这个问题应该如何解决呢

Ellohiye avatar Ellohiye

添加注意力出错

3 comment

作者你好！

haihaiwolai avatar haihaiwolai

关于模型参数问题

4 comment

![image](https://github.com/C1nDeRainBo0M/AGCA/assets/57579741/075bc1cf-89ed-47bc-a384-d7b725698ca6) 作者您好，看论文的时候关于这部分的参数量我有些疑问，为什么加ECA反而参数量还降低了呢？而且加上其他CBAM参数量为啥保持不变呀？

Ellohiye avatar Ellohiye

网络代码

1 comment

作者，您好，可以分享完整的网络代码吗

yfhdm avatar yfhdm

← Metadata

17
Stars
1
Forks
Watchers

Owner

verified publisher icon C1nDeRainBo0M

Metadata

The core code of AGCA: An Adaptive Graph Channel Attention Module for Steel Surface Defect Detection

Back