tsingqguo / inpaint4shadow Goto Github PK

I have looked at the method of calculating metrics psnr and ssim in your code. It seems that your method of calculating metrics is slightly different from the methods in other papers. I would like to know how your RMSE is calculated. Can you also include the RMSE calculation method in Evaluation. py

Hi, thank you for your amazing work, I wonder what's the number of iter u used for training?

Cuz I got inf and 0 loss after 10,000 iters for the inpainting pretraining already and get the error.

Hi!

I'm Thank you for your work!
Could you tell me how can I run inference using your model on the SRD dataset? It contains only images with shadow and without shadow. As I understand, your model requires self.inpaint_model.process(images, masks, shadows).
Could you explain what are images, masks, and shadows in the case of the SRD dataset?

Thank you in advance.