search

silverbottlep / meta_trackers

126 stars 41 forks source link

I have solved many bugs in the code! #4

Closed zhengzhenxian closed 6 years ago

zhengzhenxian commented 6 years ago

Great job! 1.the mat file is not imagenet-vgg-verydeep-16.mat but the imagenet-vgg-m.mat, you can download in http://www.vlfeat.org/matconvnet/models/imagenet-vgg-m.mat

  1. The cpu version doesn't work, so GPU is necessary for you!
silverbottlep commented 6 years ago

@zhengzhenxian thanks! by the way, what do you mean by mat file is not imagenet-vgg-verydeep-16.mat?

zhengzhenxian commented 6 years ago

Yeah, I have run the model using the pre-trained vgg-verydeep-16 network and I found it doesn't work so I replace the mat by magenet-vgg-m.mat(http://www.vlfeat.org/matconvnet/models/imagenet-vgg-m.mat). And it works! Thanks for your great jobs!!

silverbottlep commented 6 years ago

@zhengzhenxian I just checked, but I think it should be imagenet-vgg-verydeep-16.mat (http://www.vlfeat.org/matconvnet/models/imagenet-vgg-verydeep-16.mat)?

jimmy-dq commented 6 years ago

@silverbottlep I check that the model used in the code is imagenet-vgg-m.mat, which is also used in the MDNet.

silverbottlep commented 6 years ago

@zhengzhenxian @handworker OMG, my fault, CREST used imagenet-vgg-verydeep-16.mat, MDNet used imagenet-vgg-m.mat. I used the same networks for fair comparison. sorry for confusion, I will fix readme