search

kuangliu / pytorch-fpn

Feature Pyramid Networks in PyTorch
467 stars 123 forks source link

fpn structure doubt #3

Open dongzhuoyao opened 6 years ago

dongzhuoyao commented 6 years ago

in fpn.py file.

# Top-down
        p5 = self.toplayer(c5)
        p4 = self._upsample_add(p5, self.latlayer1(c4))
        p3 = self._upsample_add(p4, self.latlayer2(c3))
        p2 = self._upsample_add(p3, self.latlayer3(c2))
        # Smooth
        p4 = self.smooth1(p4)
        p3 = self.smooth2(p3)
        p2 = self.smooth3(p2)

should be 

# Top-down
        p5 = self.toplayer(c5)
        p4 = self._upsample_add(p5, self.latlayer1(c4))
        p4 = self.smooth1(p4)
        p3 = self._upsample_add(p4, self.latlayer2(c3))
        p3 = self.smooth2(p3)
        p2 = self._upsample_add(p3, self.latlayer3(c2))
        p2 = self.smooth3(p2)
        # Smooth

?

Superlee506 commented 6 years ago

@dongzhuoyao @kuangliu Where is "P6"? The original paper also contains "P6".

linghai06 commented 6 years ago

@dongzhuoyao no, they are not. check this: image

zzzz94 commented 5 years ago

P6 is contained in section 4.1.

Muzijiajian commented 5 years ago

@dongzhuoyao @Superlee506 @linghai06 @zzzz94 @kuangliu How can I use this model into my own classification tasks?