DanielXu123

Did you all trying convert modelo from keras to caffe?

Hi, @xevolesi . Have you tried to combine those three datasets together? I checked the code in ./parsing/dml_csr/dataset. it suppose only added CelebA-HQ-Mask for training.

Hey, @ssbilakeri , have you figured out how to solve this problem ?

@ssbilakeri All these "writer" related stuff seems being used in tensorboard. So, if u don't need tensorboard to show some results, those "writer" related codes can be removed.

@Avsynthe Same thing in Linux, added to 100GB RAM , is there any possible solutions?

@Wynneve I reinstalled torch from 2.1.0 to 2.0.1 , but it shows mu xformers cannot be activate correctly, could you please help to check what your xfomers version?

And, where should I place these files? Especially for the python script. I modified the caffe_root direction, but can't run correctly.

So, bascially, the transfered caffemodel can not get the same results as the keras .h5 model predict?

Very appreciate for your help, Eric! I download the whole latest version, and follow your instructions, but met some problem at the beginning( the build_win.cmd part). Maybe it's related to...

Hi, Eric. I have successfully ran the demo_webcam.cmd. The previous problem may caused by windows caffe. I used the version of Microsoft instead of BVLC, which bring some missing lib...