search

piotrkawa / deepfake-whisper-features

Implementation of the paper "Improved DeepFake Detection Using Whisper Features"
MIT License
72 stars 4 forks source link

Can u share the project Complete directory?? #12

Closed Noochila closed 4 months ago

Noochila commented 4 months ago

like share where we should we place pretrained model, where we should place dataset,Like share complete directory structure

piotrkawa commented 4 months ago

Hi, there are no requirements for specific locations. You can specify them in a config file (model's checkpoint path, e.g. configs/training/whisper_frontend_lcnn.yaml) and by using args of a script (e.g. train_and_test.py).

The datasets can be downloaded from official sources - linked in the README. Please note that we used another metadata file.

Noochila commented 4 months ago

Thank you