backseason / PoolNet

Code for our CVPR 2019 paper "A Simple Pooling-Based Design for Real-Time Salient Object Detection"
MIT License
629 stars 153 forks source link

About self.net.train() #78

Open yuzheyyyy opened 3 years ago

yuzheyyyy commented 3 years ago

I found that both in solver.py and joint_solver.py, when doing buid_model(), you just set self.net.eval(), but when we do train(), there is no command to set self.net.train(), which is commented in build_model().

backseason commented 3 years ago

You may refer to issues #30 & #73 for more details.