search

BookerDeWitt / MSBDN-DFF

The source code of CVPR 2020 paper "Multi-Scale Boosted Dehazing Network with Dense Feature Fusion"
334 stars 69 forks source link

Details in models #7

Open apache224 opened 3 years ago

apache224 commented 3 years ago

HI, I've just started learning machine learning for a few months，somethings I don't understand. I want to know in class "ResidualBlock",out= self.conv2(out) * 0.1, why 0.1?

BookerDeWitt commented 3 years ago

I follow the instruction in https://github.com/SeungjunNah/DeepDeblur_release to avoid the potential gradient explosion during the training stage. Hope this can help you :)

apache224 notifications@github.com 于2020年9月7日周一 下午7:38写道：

HI, I've just started learning machine learning for a few months，somethings I don't understand. I want to know in class "ResidualBlock",out= self.conv2(out) * 0.1, why 0.1?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/BookerDeWitt/MSBDN-DFF/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACAO55BFYGLEERLMGSZDEFLSETA2VANCNFSM4Q6AZ5HA .

apache224 commented 3 years ago

Thanks :)