zhangzc21 / petsgan Goto Github PK

Hi! I would like to train your model on a 4500x2255 size image: what is the max output size I can get? I tried to set --max_size to 2000 but generated output images in "final_syntheses" folder are 256x128. I'm working with a RTX 5000 16 GB.

python main.py --input_dir Input --image_name img.png --max_size 2000 --gan_epoch 5000

Thank you!

Hi, I have question about refined result. After train is done, there are results named "hr{num}" and "hr_refined{num}".
I guess the difference between these two result is TTSR, which is reference-based image super-resolution model.
I can't find any explaination about this from your paper. So if you don't mind could you explain about this?
And i have two explicit question. Is it possible to use TTSR 1x scale?(output size is same with input) and is reference image used for TTSR is G.T. without rescaling?

I'd appreciate it if you could leave a comment.

I'm quite interested in following your work after reading your paper. But I ran into some issues.

Where can I get datasets in paper ?

How to replicate your quantified result in paper?

I will be appreciated for your reply, thank you very much!
Best wishes

When I clone this project and try to reproduce it, I found that I could not find the vgg file that can be imported from Pretrained directory, so the inversion.py file cannot work as I wish, whereas I have followed the README guidance and downloading pertained files and put them into Pretrained directory. Did the corresponding file be removed to another place or it was missing somehow? Would it be fixed in the near future?
BTW, would there be any instructions in reproducing this project?
Thanks a lot.

Hello.
Thank you for your great work that evolved single image generation model !

Then, I want to try your model to image manipulation such as style-transfer, but I think there is only code of main training.

So, could you instruct me a way that run image manipulation tasks, like main_training

python main.py --input_dir xxx --image_name xxx.jpg

Sorry to bother you, but I would be happy to receive your reply.

Hi i'm interested in your research. But i can't find appendix for this paper at AAAI and arXiv page.
Is there any ways to get appendix for this paper?