mrrobot2211 / pytorch-deep-photo-enhancer Goto Github PK

Thank you for your code. It is very clear and helpful! I'm trying to implement this work on my own data. I'm a little bit confused about is it necessary to pre-train weights for the generator? The instructions in README.md let me think I need to do SL before doing UL. But it isn't mentioned in the paper. Is a pretrained generator necessary?

Thank you!

I run command

!python unbounded_inference.py --generator_model 'gan2_train_28_40.pth' --input_image 'teaser.jpg' --output_image 'xui.jpeg'

But don't see any different in comparison with original.
Result

Original

Hi, MrRobot. When I'm trying to run 'python 2WayGAN_Train_v3', I am getting this error:
File "2WayGAN_Train_v3.py", line 9, in
from libs.networks import *
File "/data/glusterfs_cv_04/11121734/pytorch-deep-photo-enhancer-master/libs/networks.py", line 61, in
from util.image_pool import ImagePool
ModuleNotFoundError: No module named 'util'
but I can't find where the util.py is. Looking forward to your reply, thank you very much.

Hii Mr Robot , when i am trying to resize my image in resize.py ,i am getting this error
PermissionError Traceback (most recent call last)
in
49 # resize_image("/home/felipe/deep-photo-enhancer-master/images_LR/Expert-C/Testing")
50 # resize_image("/home/felipe/deep-photo-enhancer-master/images_LR/Expert-C/Training1")
---> 51 rename_MIT_files("/Users/HP/Desktop/pytorch-deep-photo-enhancer/images_LR/Expert-C/Training2")
52 resize_image("/Users/HP/Desktop/pytorch-deep-photo-enhancer/images_LR/Expert-C/Training2")
53 # resize_image("/home/felipe/deep-photo-enhancer-master/images_LR/Expert-C/Training2/hdrs")

in rename_MIT_files(dir)
24 os. remove(subdir+'/'+file)
25 else:
---> 26 os.rename(subdir+'/'+file,subdir+'/'+file.split(".")[0]+".tif")
27
28

PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: '/Users/HP/Desktop/pytorch-deep-photo-enhancer/images_LR/Expert-C/Training2\1/bad.tif' -> '/Users/HP/Desktop/pytorch-deep-photo-enhancer/images_LR/Expert-C/Training2\1/bad.tif'

Hi there, have you tried this code on DPE's unpaired data split and achieved the same result?