Open bill1969 opened 2 years ago
had the same problem following Installation guide
The solution is I give you the folder structure,you need to have "sync_batchnorm" folder inside
"Bringing-Old-Photos-Back-to-Life\Face_Enhancement\models\networks" folder
so the final structure is going to be like in the picture.Make sure you don't overwrite any file.
I tell you what guys? After stage 1 it is not worth to do the rest of the stages.Because it is using dlib for faces and even if you compile it with cuda to run faster,it takes time to process the faces.What is worst is that the final results is not worth the time and processing power spent on it.
So I suggest you to edit the run.py file and comment out stage 2,3,4 for efficiency.
These are the original files :
These 2 are without face enhancement :
These 2 are with face enhancement :
For face enhancement you better use gfpgan or gpen.
For convenience edited run.py file: run.zip
I have recorded 2 videos about this project is being used if you wanna watch:
https://www.youtube.com/watch?v=PK3Qpyld9fQ&t
https://www.youtube.com/watch?v=2gS0wzvWZuA&t
All credits goes to related project developers.
Running Stage 3: Face Enhancement Traceback (most recent call last): File "test_face.py", line 15, in
opt = TestOptions().parse()
File "E:\picture_fix\Bringing-Old-Photos-Back-to-Life\Face_Enhancement\options\base_options.py", line 262, in parse
opt = self.gather_options()
File "E:\picture_fix\Bringing-Old-Photos-Back-to-Life\Face_Enhancement\options\base_options.py", line 197, in gather_options
parser = model_option_setter(parser, self.isTrain)
File "E:\picture_fix\Bringing-Old-Photos-Back-to-Life\Face_Enhancement\models\pix2pix_model.py", line 12, in modify_commandline_options
networks.modify_commandline_options(parser, is_train)
AttributeError: module 'models.networks' has no attribute 'modify_commandline_options'
Finish Stage 3 ...