ludakreiss/deepfake-detection-under-compression-and-resizing ? reverse-engineered prompt

Build me a simple Python project for working with the FaceForensics++ dataset for deepfake detection experiments.

I want a clean setup that helps me download the dataset with the official FaceForensics++ download script, organize the files, and make it easy to run experiments where I compare detection results on different compression levels and resized versions of the videos. If it makes sense, add a few small helper scripts for downloading, checking the dataset, and preparing data folders so I can get started quickly.

Please keep it practical and easy to run from the command line, with clear instructions in the README for setup and usage. If you need to look up the current FaceForensics++ download instructions or the latest official docs online, go ahead and do that.