The code for our newly accepted paper in Pattern Recognition 2020: "U^2-Net: Going Deeper with Nested U-Structure for Salient Object Detection."
8.52k
stars
1.47k
forks
source link
BCELoss was unstable so just get logits from model and use BCEwithLog… #309
Open
deshwalmahesh opened 2 years ago
Changes according to Performance Tuning: Official docs
bias = False
in Line 9 of model ofu2net.py
torch.backends.cudnn.benchmark = True
and use Mixed Precision TrainingBceWithLogitLoss
: Line 29 in place ofBCELoss
of `trainingd0-d6
in place ofSigmoids
: Line 525 ofu2net.py
torch.backends.cudnn.benchmark = True
: Line 69 of training