Confusion about the self.max_of_8_tensor() operation

XinJCheng / CSPN

Convolutional Spatial Propagation Network

498 stars 93 forks source link

Confusion about the self.max_of_8_tensor() operation #11

Closed DecaYale closed 5 years ago

DecaYale commented 5 years ago

Hi, I noticed that you employ a function self.max_of_8_tensor(...) to obtain the final results. Why? It seems hard for me to associate the code with the operation proposed in the paper.