why there are attentions in ResNet?

arthurdouillard / CVPR2021_PLOP

Official code of CVPR 2021's PLOP: Learning without Forgetting for Continual Semantic Segmentation

https://arxiv.org/abs/2011.11390

MIT License

145 stars 23 forks source link

Closed zhaoedf closed 2 years ago

zhaoedf commented 2 years ago

why? at first i thought this might be related to inplace_abn, then i checked the paper of inplace_abn, found no relevant info.

arthurdouillard commented 2 years ago

What I call attentions there is simply the ultimate feature maps of a resnet block (mod*) just before the ReLU.

For here x = ReLU(att).

zhaoedf commented 2 years ago

What I call attentions there is simply the ultimate feature maps of a resnet block (mod*) just before the ReLU.

For here x = ReLU(att).

oh i understood, thanks