Fausto Milletari

Hello, I think you should use 3d caffe which is available in another repository on my github. Latest cudnn should work. Fausto Milletarì Sent from my iPhone > On 19....

It sounds a bit weird because caffe is complaining that he does not know what a python layer is. But python layers are supported in caffe since long time! And...

when you compiled caffe you did also make pycaffe and then make install, right? > On 19 Apr 2017, at 10:44, wxde wrote: > > I have check my pythonpath...

You can use the tensorboard. Point it to the directory where you save the models. There should be summaries there. Fausto Milletarì Sent from my iPhone > On 20. Jun...

the answers are: yes. (it’s just a trick to speed things up a little. it pushes in the right direction both according to background and foreground.) yes. (minimising -dice) >...

in caffe one must write own derivatives. therefore if you take the derivative of dice and you add a minus in front of it you get the derivative of -dice....

Actually for the background I'm just trying to make it do the opposite of what the foreground is doing, intuitively. One even better thing to do could be to write...

Apparently there is some Nan or something like that. Check the data that is fed to the network. Be sure that there is some voxels that are part of foreground....

I don’t think so. one thing you can do is to edit pylayer.py and put an assertion that makes sure that every ground truth label volume has at least one...

unfortunately i don’t know how to make caffe work on windows. > On 17 Nov 2016, at 17:03, EhsanSHV [email protected] wrote: > > Hello, faustomilletari > Thank you for sharing...