Question about p1dataset.py

[clovaai]

Question about p1dataset.py

[8uos]

Hi.

The model is only available to generate the components in the style characters in phase 1 and so the style characters should contain all the components.
To satisfy this, we randomly picked a character that contains the component for each component of the target character. (please see the class function sample_char.)

In this case, extracting the remaining components of each style characters only wastes the memory; because they are not used for the generation.
So this code excludes the remaining components and only returns the target components.

[yjunej]

I get the picture. Thank you