Benchmark-ready packaging of the DeepFake (DF) evaluation partition from ASVspoof 2021 for speech anti-spoofing and synthetic / deepfake voice detection.
This dataset contains the DF evaluation subset of the ASVspoof 2021 challenge. The task is binary classification: bonafide (genuine human speech) vs. spoof (synthetic, converted, or otherwise manipulated speech). The original dataset is available at… See the full description on the dataset page:
https://huggingface.co/datasets/SpeechAntiSpoofingBenchmarks/ASVspoof2021_DF.
