how to train with VGG16?

hbilen / WSDDN

Weakly Supervised Deep Detection Networks (CVPR 2016)

148 stars 53 forks source link

Closed yuxinyan closed 7 years ago

yuxinyan commented 7 years ago

i have downloaded vgg16 model for matconvnet and chage the model path to it. I have set vggdeep = 1 in prepare_wsddn.m. What else should i do?

hbilen commented 7 years ago

Just setting opts.modelPath to vgg16 should work without requiring any further change.

yuxinyan commented 7 years ago

@hbilen Thanks a lot