A personal implement of CAM(class activation mapping) part of CVPR 2018 Revisiting Dilated Convolution: A Simple Approach for Weakly- and Semi- Supervised Sematic Segmentation.
LeNet/VGG16+GAP+CAM on CIFAR10.
Not finished yet...
- python 3.6.8
- pytorch 1.0.1
- cudnn 7.6.0
- numpy 1.16.4
python main.py
If you want to use the version of the GPU, please modify the code at the comment 'GPU'.
The first line is the results of training 1/200 epoch, and the second line is the results of training 21/200 epoch.