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

Reproducing COCO-stuff #40

Open sigeisler opened 3 years ago

sigeisler commented 3 years ago

Hi,

thanks for sharing this great code base including pre-trained models! However, I would like to train the COCO-stuff model from scratch. Can you give me some instructions on what to do? I just want to make sure that I do not miss sth.

Thanks, Simon

zsyzzsoft commented 3 years ago

In addition to our pix2pix mode, you need to add an embedding layer to the input label map. When processing data, also need to make sure that the input label map is correctly resamped (using nearest interpolation)