why use maxpooling in p1p2p3?

GNAYUOHZ / ReID-MGN

Simple pytorch unofficial implement of paper: Learning Discriminative Features with Multiple Granularities for Person Re-Identification

235 stars 61 forks source link

why use maxpooling in p1p2p3? #25

Open zhishao opened 4 years ago

zhishao commented 4 years ago

In network.py you using maxpooling in branches, not avgpooling in the paper, why ?

GNAYUOHZ commented 4 years ago

In some details, it is different from the original paper, such as dimensions, such as optimizer. I can't get good results if I follow the original paper completely