Details on training and testing

davide-coccomini / Combining-EfficientNet-and-Vision-Transformers-for-Video-Deepfake-Detection

Code for Video Deepfake Detection model from "Combining EfficientNet and Vision Transformers for Video Deepfake Detection" presented at ICIAP 2021.

https://dl.acm.org/doi/abs/10.1007/978-3-031-06433-3_19

MIT License

239 stars 60 forks source link

Details on training and testing #41

Open Ankit2527 opened 1 year ago

Ankit2527 commented 1 year ago

Hi @davide-coccomini, As you mentioned in the paper, that you have used both FF++ and DFDC dataset to train the model:

Which version of FF++ dataset do you use (raw/c23/c40)?
As for DFDC only the training data is available and the public and private test sets are not released. How do you use the DFDC dataset? Do you do a split from the training set itself into train/val/test?

davide-coccomini commented 1 year ago

Hi, the FF++ version used is c23. The DFDC dataset, labels and splits can be found at: https://dfdc.ai/

Ankit2527 commented 1 year ago

Thank you @davide-coccomini for your quick response!

viethc02 commented 1 year ago

I got some errors when trying to download dfdc data. Have you downloaded it yet? I will be very grateful if you can share with me the test dataset via my email. Thank you so much.

davide-coccomini commented 1 year ago

Hi, what kind of errors are you getting?

viethc02 commented 1 year ago

Exceeded daily email limit for the operation or the account. If a higher limit is required, please configure your user pool to use your own Amazon SES configuration for sending email.

Thanks for replying! I got that error when trying to register an account on dfdc.ai. I tried to fix it but still can't sign up and some of my friends said it was due to website error. Do you have any suggestions for me?

davide-coccomini commented 11 months ago

Unfortunately not, many people are having issues with DFDC dataset downloading but we can't do anything...