michel-liu / fatformer Goto Github PK

I have a question, because you don't have open source yet. After reading the paper, I don't understand, when detecting whether an image is a synthetic, only the image needs to be input, or the corresponding text of the image is needed?

Hello,
Your paper is truly impressive. However, the structures of both the FAA module and LGA module are rather intricate. I have already reached out to you via email, and I was wondering if it would be possible for you to provide a simplified version of the implementations for these two modules. Your assistance would be greatly appreciated.

Hello, thank you for your excellent work! I have a question, since FatFormer is built based on the CLIP framework, what text encoder is used in Table 4 for the Swin transformer and MAE. If the CLIP text encoder is used here, the pre-trained features of images and text prompts are not initially aligned.