SpeechAntiSpoofingBenchmarks/ASVspoof2021_LA
ASVspoof 2021 LA Benchmark-ready packaging of the Logical Access (LA) evaluation partition from ASVspoof 2021 for speech anti-spoofing and synthetic / deepfake voice detection. Overview This dataset contains the LA evaluation subset of the ASVspoof 2021 challenge. The task is binary classification: bonafide (genuine human speech) vs. spoof (synthetic, converted, or replayed speech). The LA partition adds realistic telephony codec / transmission conditions on top… See the full description on the dataset page: https://huggingface.co/datasets/SpeechAntiSpoofingBenchmarks/ASVspoof2021_LA.
Add RawBMamba submission for ASVspoof2021_LA (EER=9.37567743334056) (#20)
Add AMSDF submission for ASVspoof2021_LA (EER=7.83112941686538) (#19)
Add SSL Feature Gating submission for ASVspoof2021_LA (EER=6.492521135920224) (#18)
Add CLAD submission for ASVspoof2021_LA (EER=12.372642531974854) (#17)
Add DF Arena 500M submission (eer=5.7781%, scoring-verified)
Add DF Arena 1B submission (eer=4.8071%, scoring-verified)
Add maintainer contact: email + Telegram channel
Deprecate old contact email; add kborodin.research@gmail.com
Add Spectra-AASIST submission: ASVspoof2021_LA (#15)
Submit Spectra-0 on ASVspoof2021_LA (EER 6.58%) (#14)
Add Spectra-AASIST3 submission (ASVspoof2021_LA) (#13)
Submit AASIST3 results on ASVspoof2021_LA (EER 32.06%) (#12)
Add XLSR-SLS submission (ASVspoof2021_LA, EER 7.39%) (#11)
Add Nes2Net submission (ASVspoof2021_LA) (#10)
W2V2-AASIST: fill reproduction block (scoring)
Add W2V2-AASIST submission (#9)
Add Whisper-MFCC-MesoNet submission (#8)
Add RawTFNet submission (ASVspoof2021_LA) (#6)
Add AASIST-L submission (#5)
Add AASIST submission (#4)
Upload submissions/rescapsguard.yaml with huggingface_hub (#3)
Add Res2TCNGuard submission (EER 13.67%, scoring-reproduced) (#2)
Add random-baseline submission (ASVspoof2021_LA) (#1)
Add ASVspoof2021_LA canonical parquet build (181,566 LA eval trials)
initial commit