search

Xiaoqi-Zhao-DLUT / GateNet-RGB-Saliency

(IJCV 2024&ECCV 2020 Oral) Suppress and Balance: A Simple Gated Network for Salient Object Detection
MIT License
162 stars 15 forks source link

I retrain the model, however the prediction map is meaningless map. What mighe be the error? #3

Closed peipei2015 closed 3 years ago

peipei2015 commented 3 years ago

A part of the training log:

[iter 99500], [total loss 0.04293],[loss1 0.02237],[loss1 0.02055],[lr 0.0000085085186] [iter 99600], [total loss 0.04290],[loss1 0.02238],[loss1 0.02052],[lr 0.0000069635378] [iter 99700], [total loss 0.04236],[loss1 0.02209],[loss1 0.02027],[lr 0.0000053791028] [iter 99800], [total loss 0.04294],[loss1 0.02237],[loss1 0.02057],[lr 0.0000037400418] [iter 99900], [total loss 0.04292],[loss1 0.02237],[loss1 0.02055],[lr 0.0000020132107] [iter 100000], [total loss 0.04314],[loss1 0.02248],[loss1 0.02066],[lr 0.0000000316228]

Xiaoqi-Zhao-DLUT commented 3 years ago

@peipei2015 I guess you may forget to apply the sigmoid fuction for the prediction.

peipei2015 commented 3 years ago

You uncomment the sigmoid operation in test code.

peipei2015 commented 3 years ago

Right, thank you.