DIC Model

[cam207]

I wanted to ask do i have to download the DIC model run it using the images of the dataset then use the results for the vision transformer or is everything already in the current repo and i can just directly run it without needing to do anything else?
Thank you

[hbin0701]

Hi, as stated in README.md, you'll need to:

1. First downlaod images from FER+.
2. We have used DIC model for SR module of our paper, so apply super-resolution (x4) to each image with the pretrained model of DIC_HELEN.pth.
3. Finally, put the SR images accordingly in the /data/train and /data/test directory.

---

... have to download the DIC model run it using the images of the dataset then ... "
So yes. this would be the case.