Realtime_Multi-Person_Pose_Estimation icon indicating copy to clipboard operation
Realtime_Multi-Person_Pose_Estimation copied to clipboard

Published 20 hours ago • verified publisher icon ZheC

Why only first 10 layers of VGG19 ?

Open GajjarMihir opened this issue 6 years ago • 0 comments

Why are we using only first 10 layers of VGG19? We can use maybe a few less layers or a few more layers? I guess the reason would be that using the first 10 layers or all the 19 layers would be having almost the same accuracy. Any help would be really appreciated.

GajjarMihir avatar May 05 '18 08:05 GajjarMihir