How important is dropout in DAN?

shenyunhang / DRN-WSOD-pytorch

Enabling Deep Residual Networks for Weakly Supervised Object Detection

Apache License 2.0

50 stars 10 forks source link

Open vadimkantorov opened 3 years ago

vadimkantorov commented 3 years ago

Droput prob here is quite high - 0.5, and it is not discussed in the paper...

shenyunhang commented 2 years ago

Sorry for my late response.

I think it should be important at least for classification objective in WSOD.