Hi @MhLiao
I am successfully able to train TextBoxes on my custom dataset. But while training, I am not able to see the mbox loss in the validation (or test) phase.
On parsing the file VGG_text_longer_conv_400x100.log using parse_log.py (utility function in caffe), I am getting NumIters, Seconds,learning rate, mbox loss in the VGG_text_longer_conv_400x100.log.train and NumIters, Seconds, Learning, detection_eval in VGG_text_longer_conv_400x100.log.test. Currently I am checking that the detection_eval (mean average precision) should increase with the increase in the number iterations. Am I doing the right thing?
How can I get mbox loss in the validation (or test) phase? How can I figure out whether I am training the model correctly?
Hi @MhLiao I am successfully able to train TextBoxes on my custom dataset. But while training, I am not able to see the mbox loss in the validation (or test) phase.
On parsing the file VGG_text_longer_conv_400x100.log using parse_log.py (utility function in caffe), I am getting NumIters, Seconds,learning rate, mbox loss in the VGG_text_longer_conv_400x100.log.train and NumIters, Seconds, Learning, detection_eval in VGG_text_longer_conv_400x100.log.test. Currently I am checking that the detection_eval (mean average precision) should increase with the increase in the number iterations. Am I doing the right thing?
How can I get mbox loss in the validation (or test) phase? How can I figure out whether I am training the model correctly?