Tamar Rott Shaham
Tamar Rott Shaham
the outputs are automatically saved in the output folder
Hi, Thanks! I'm not sure I fully understand the question: Do you want the generated images to be similar to the training image at some fixed position? You can maybe...
Thanks, this function indeed can only work with `gen_start_scale=0`. This because when starting the generation from a finer scale, you need to also input a specific image to this scale....
1. as we explain in the paper we combine the generated image with the original background. We do this using a binary mask. 2. the way to use the mask...
pytorch 1.4 should solve this
In the current version of the code, the harmonized image should have the same size of the original image. z_curr is the current noise input sample. in order to work...
you can create the naively pasted version and the mask manually, using photoshop/paint.net or similar programs
you should change nc_im to 1
You should change --nc_im to 1 instead of 3. Pay attention that you will need to have large GPU memory in order to train on 2048x2048 image.
once you have a trained SinGAN model, it is subtitle for generation of the image you trained it for. For editing and harmonization you can input different reference images (as...