Help needed in getting vgg16 to train.

[yxc...@gmail.com]

Hi, 

I have attached my model.prototxt in this post. Right now, my model could not converge and its loss is jumping erratically. 

Since I am relatively inexperienced, I wonder what I should change in my model to get it to work.

I understand that the original model does not have batch norm but i think batch norm should help it to converge and should not be the main problem.

[akashgu...@gmail.com]

Hey,

Were you able to solve this problem?