I don't know how to start about stylenet HOT 2 CLOSED

Data loader and preprocessing depend on the data, flickrstyle10k dataset.
However, it is not open, so I can't share the dataset.
I asked authors to share it when I implemented this.
If you use another data, you may have to modify data loader.

In my data, a line in factual_train.txt includes image name, id of the caption for the same image and caption.
ex) 111111.jpg#0 this is a caption.
a line in funny_train.txt includes only caption with humorous style.

In addition, my code is not completed.
It can generate sentences, but they are not related to images.
I consider image features are not transmitted properly because of my implementation mistakes.
However, now I am busy so don't have enough time to find bug and modify it.
I am sorry for it.

from stylenet.

Thanks a lot for your reply~

from stylenet.