ocr-vqgan
ocr-vqgan copied to clipboard
`resolution` and `size` in train config
Hello @joanrod , thank you for your amazing work!
I wonder resolution
under model's ddconfig
and size
under data
in config yaml file should be same or not.
In the original Taming-Transformer repo, both are set to 256.
But in this repo, the image size is set to 384.
Is there any reason that they are set to the different numbers?
Thanks