search

meteorshowers / RCF-pytorch

Richer Convolutional Features for Edge Detection model in pytorch CVPR2017
509 stars 145 forks source link

lb[np.logical_and(lb>0, lb<128)] = 2 #45

Open LunarSun opened 4 years ago

LunarSun commented 4 years ago

why lb[np.logical_and(lb>0, lb<128)] = 2? in data_loader.py { class BSDS_RCFLoader(data.Dataset):/def getitem(self, index):}

LunarSun commented 4 years ago

I got it .it is used to mask the weight in function.py /def cross_entropy_loss_RCF(prediction, label):