search

marcellacornia / mlnet

A Deep Multi-Level Network for Saliency Prediction. ICPR 2016
MIT License
94 stars 37 forks source link

some question about training process。 #21

Open YanDingXin opened 4 years ago

YanDingXin commented 4 years ago

When I used your model to train our dataset, the loss function changed to nan over time.Do you know why?