As stated in the guideline, the FaceForensics++ dataset can be freely used for model training and internal testing (refer to the guideline and the GitHub link below).

FaceForensics++ dataset (Github): https://github.com/ondyari/FaceForensics
(**Private test data used for final evaluation has not been released.**)

---
You can use the public test dataset to verify that your final Kaggle notebook runs correctly and that the output format is appropriate for submission.

#### 📂 Public_test_dataset
- This folder contains video files for public testing purposes.

#### 📄 Metadata.json
- JSON file that contains metadata for each video, such as the label (REAL or FAKE), the original source video, and the manipulation method (e.g., face swap).

#### 📄 Sample_submission.csv
- Sample file demonstrating the required format for predicted label output.