Skip to first unread message
el
Dec 18, 2015, 7:34:39 AM12/18/15
to Caffe Users
Hello..
I want to finetune a network to own public dataset which only provides train and val set. My question is whether this val set is used in both validating and testing. In my train_val.prototxt, where I point out include {phase:TEST and TRAIN} I use train set as source to TRAIN phase and val set in TEST phase, and then I run caffe train --solver=....
Does caffe cross-validate through train set or it uses val set to do that?
Most of the public datasets do not include test sets.
In testing phase, I have just to run caffe test --model=.... with the same train_val.prototxt file, or should I change the TEST source...
Thank you in advance :)
I want to finetune a network to own public dataset which only provides train and val set. My question is whether this val set is used in both validating and testing. In my train_val.prototxt, where I point out include {phase:TEST and TRAIN} I use train set as source to TRAIN phase and val set in TEST phase, and then I run caffe train --solver=....
Does caffe cross-validate through train set or it uses val set to do that?
Most of the public datasets do not include test sets.
In testing phase, I have just to run caffe test --model=.... with the same train_val.prototxt file, or should I change the TEST source...
Thank you in advance :)
Reply all
Reply to author
Forward
0 new messages