search

zsyzzsoft / co-mod-gan

[ICLR 2021, Spotlight] Large Scale Image Completion via Co-Modulated Generative Adversarial Networks
Other
444 stars 67 forks source link

Current state of Pix2Pix? #18

Open richardburleigh opened 3 years ago

richardburleigh commented 3 years ago

Amazing work on this project!

I see quite a few references to Pix2Pix in the code and wonder how much has been implemented so far?

Out of curiosity I wanted to add the ability to load pre-generated masks of objects, such as glasses, to see the impact on the results, but it seems a lot of this logic has been built already.

zsyzzsoft commented 3 years ago

The pix2pix option is for image-to-image translation, and currently it is implemented on the network side but not yet on the data side. I think it may not be useful in your case?