Cannot Get Decent Results in Unseen Images

[SaidPanoculum]

Hi,

I am testing unseen images on the pretrained model daclip_ViT-B-32.pt and it seems that model is not working well. There is no or very slight changes in the images. As an example, here is the original image and the output image, respectively.

Is there a problem related to the pretrained model or is this a normal behavior on unseen images due to the limitations of the training dataset of the pretrained model?

Thanks.

[Algolzw]

Hi it's a normal result due to the dataset limitations (we didn't add the old photo recovering task in our training). For the above case, you can try this work.