lengstrom
Checkpoint Files
I have been using https://github.com/hwalsuklee/tensorflow-fast-style-transfer to do some hobby style transfering. There, I train the models and come up with a .ckpt file that gets used; I am unable to use that file for this off the shelf. Is anyone aware of an easy fix so that I do not have to retrain models for hours on end?
The models for each repo are inherently different. It would take insane amounts of time to manually "fix" the files to be interchangeable. If you're looking for a neural style that is immensely faster, try a method that uses metalearning, such as my code here: https://github.com/JEF1056/Reconstruction-Style
Is it faster than this one? Because this one seems really solid.
Edit: I see the one you linked doesn't work on windows.
@spot92 https://colab.research.google.com/ will help if you don't have linux
And yes, that repo is significantly faster, with training times that can take at most 3 minutes on GPU