search

yoctta / multiple-attention

The code of multi-attention deepfake detection
241 stars 54 forks source link

Running evaluation on FF++,DFDC,Celeb-DF #25

Closed sharadchandakacherla closed 1 year ago

sharadchandakacherla commented 1 year ago

@yoctta please help us out,

  1. I want to run the code in evaluation mode, however there is no celeb.json , i'm guessing this is for Celeb-DF-v1. Can you please outline steps how to evaluate the code?
  2. Also could you please the exact file structure of your dataset directory and how to populate it?
  3. There is no folder called runs with any config.pkl file.

P.S. : I just wan't to reload the shared weights and run inference on the test sets

Thanks

sharadchandakacherla commented 1 year ago

@yoctta So i made dirs and trainconfig to generate the config.pkl file, and then was able to load the weights into EfficientNet-b4 from the .pth , but not sure how to get the preprocessing running as im not sure how to get retina face running, should I use the pip install retina-face or use https://github.com/biubug6/Pytorch_Retinaface?

sharadchandakacherla commented 1 year ago

closing the issue, was able to run evaluation successfully.

DatDaiGia commented 1 year ago

@sharadchandakacherla Could you please share how you ran the evaluation successfully? I am getting stuck in loading pre-trained weights. Many thanks in advance!