Is it true that you didn't train the parameters of VGG16?

[Mistariano]

Hi,

Thank you very much for the brilliant work. :)

I just wonder why you set the parameters of VGG16 as constants:

holy-edge/hed/models/vgg16.py

Line 163 in 98cf525

return tf.constant(self.data_dict[name][0], name="filter")

According to the paper, the objective function minimizes both W and w, but it seems like you only minimized w

[mengxingkong]

i have the same question, had you resolved it?

[sandhawalia]

Initially to train faster we didn't fine-tune the Convnet backbone. Initially we just train the HED head. One can fine-tune the model one the deconv layers are trained.