About image translation

[EthanG97]

I see you mentioned the model is applicable to image translation in the paper. How to implement it?
Thanks

[jychoi118]

Hi, you just need to provide a source image as a input reference image

[EthanG97]

Just for making sure, so I need to prepare the source image and run the ilvr_sample.py as you showed as an example, right?

[jychoi118]

Yes, that's correct

[shahdghorsi]

Just for making sure, so I need to prepare the source image and run the ilvr_sample.py as you showed as an example, right?

Hello @EthanG97 were you able to run the sampler? each time I run it runs for hours without iterating or generating any output :(

[jychoi118]

Hello @shahdghorsi , did you place your input images into a folder ref_imgs/?
Or did you set PYTHONPATH variable before running the script?

[shahdghorsi]

Hello @shahdghorsi , did you place your input images into a folder ref_imgs/? Or did you set PYTHONPATH variable before running the script?
Hi, yes I did all of these. I actually left the ref_imgs/ as it is initially just to text with the given images under the folders. cat, face and portrait.

The output is as follows:

[shahdghorsi]

It turned out that the reason behind this is the big "diffusion_steps" if anyone is using a slow GPU it might take hours to generate