STIL how to train on the other datasets.

wizyoung / STIL-DeepFake-Video-Detection

Code for paper STIL: Spatiotemporal Inconsistency Learning for DeepFake Video Detection

1 stars 0 forks source link

STIL how to train on the other datasets. #1

Open ats4869 opened 8 months ago

ats4869 commented 8 months ago

Dear Authors,

thanks for making this awesome work on deepfake detection opensource. I have trained and tested on the FF++ data set, but I want to train and test on the celeb-DF-v2 data set. How should I modify your code? Do I need to reuse MTCNN to extract faces?

YuXiaolu801 commented 2 months ago

hello! have you solved the problems?