Open intervolga-school opened 2 years ago
Here you do adaptive pooling with different output sizes https://github.com/QLYoo/LFPNet/blob/master/model.py#L276 Could you please clarify why you use sizes 1 & 2 two times? Is it an error or done on purpose?
We increase the ratio of large-scale poolings to force the network to learn long range features (mainly average colors).
Here you do adaptive pooling with different output sizes https://github.com/QLYoo/LFPNet/blob/master/model.py#L276 Could you please clarify why you use sizes 1 & 2 two times? Is it an error or done on purpose?