Skip to first unread message
akashgu...@gmail.com
Feb 3, 2018, 5:37:48 PM2/3/18
to Caffe Users
Hi,
I'm new to caffe and wanted to train VGG16 face descriptor from scratch. My training loss is between 7-8 for about 15k iteration when using batch size of 64 images using standard solver values.
Was anyone able to do that and could help me with it?
Reply all
Reply to author
Forward
0 new messages