Wav2Vec2 Large AntiDeepfake
Bit-exact copy of nii-yamagishilab/wav2vec-large-anti-deepfake (the AntiDeepfake collection, nii-yamagishilab), hosted for the Speech Anti-Spoofing Arena. Trained with RawBoost data augmentation (default).
Model source
- Kind
- Other
- File size
- 1.18 GB
- Base model
- nii-yamagishilab/wav2vec-large-anti-deepfake
- Version
- vRevision 522dbc6
- Updated
- Oct 25, 2026
- Downloads
- 50
- Source
- Hugging Face
Source description
Wav2Vec2-Large-AntiDeepfake
Bit-exact copy of nii-yamagishilab/wav2vec-large-anti-deepfake (the AntiDeepfake
collection, nii-yamagishilab), hosted for the Speech Anti-Spoofing Arena.
Trained with RawBoost data augmentation (default).
- Paper: Post-training for Deepfake Speech Detection, arXiv:2506.21090
- Original code: https://github.com/nii-yamagishilab/AntiDeepfake
- Backbone: wav2vec 2.0 Large; head = AdaptiveAvgPool1d + Linear(D, 2). Score = real logit.
- License: CC BY-NC-SA 4.0 (non-commercial) — inherited from the source.
⚠️ Evaluation restriction
It is STRICTLY PROHIBITED to evaluate this model on its training / in-domain corpora: ASVspoof2019 LA, ASVspoof2021 LA/DF, ASVspoof5, CFAD, DECRO, DFADD, "Diffuse or Confuse", DiffSSD, DSD, HABLA, MLAAD, SpoofCeleb, CVoiceFake, any LibriTTS-based set (incl. SONAR, CD-ADD, LibriSeVoc), any VoxCeleb-based set, WaveFake, any FLEURS-based set, CodecFake (any variant), any AIShell3-based set, and any MLS-based set. On the Arena it is evaluated only on ArAD and In-the-Wild.
All credit to the original authors (Ge, Wang, et al., NII).
Arena results
EER% 0.95 on XMAD EER% 0.60 on PyAra EER% 6.92 on ODSS EER% 0.13 on J-SPAW_LA EER% 7.64 on DeepVoice EER% 2.89 on InTheWild EER% 5.99 on ArAD 1-SRR% 0.06 on LRLspoof EER% 0.77 on EmoFake_test 1-SRR% 6.28 on EmoSpoofTTS arena tier arena rank
Maintainer
Maintained by Kirill Borodin (SpeechAntiSpoofingBenchmarks).
- Email: [email protected]
- Telegram: @korallll_ai
Sources
1 source
Sources
1 sourceVerified Oct 25
Model artifacts
2 artifacts
Model artifacts
2 artifactsSource excerpts
3 excerpts
Source excerpts
3 excerpts--- license: cc-by-nc-sa-4.0 base_model: - nii-yamagishilab/wav2vec-large-anti-deepfake tags: - anti-spoofing - audio-deepfake-detection - speech pipeline_tag: audio-classification --- # Wav2Vec2-Large-AntiDeepfake Bit-exact copy of [`nii-yamagishilab/wav2vec-large-anti-deepfake`](https://huggingface.co/nii-yamagishilab/wav2vec-large-anti-deepfake) (the AntiDeepfake collection, nii-yamagishilab), hosted for the Speech Anti-Spoofing **Arena**. Trained with RawBoost data augmentation (default). - **Paper:** Post-training for Deepfake Speech Detection, [arXiv:2506.21090](https://arxiv.org/abs/2506.21090) - **Original code:** https://github.com/nii-yamagishilab/AntiDeepfake - **Backbone:** wav2vec 2.0 Large; head = AdaptiveAvgPool1d + Linear(D, 2). Score = real logit. - **License:** CC BY-NC-SA 4.0 (non-commercial) — inherited from the source. ## ⚠️ Evaluation restriction It is **STRICTLY PROHIBITED** to evaluate this model on its training / in-domain corpora: ASVspoof2019 LA, ASVspoof2021 LA/DF, ASVspoof5, CFAD, DECRO, DFADD, "Diffuse or Confuse", DiffSSD, DSD, HABLA, MLAAD, SpoofCeleb, CVoiceFake, any LibriTTS-based set (incl. SONAR, CD-ADD, LibriSeVoc), any VoxCeleb-based set, WaveFake, any FLEURS-based set, CodecFake (any variant), any AIShell3-based set, and any MLS-based set. On the Arena it is evaluated **only** on ArAD and In-the-Wild. All credit to the original authors (Ge, Wang, et al., NII). ## Arena results <!-- arena-badges --> [](https://huggingface.co/spaces/SpeechAntiSpoofingBenchmarks/SpeechAntiSpoofingArena?system=wav2vec2-large-antideepfake) [](https://huggingface.co/spaces/SpeechAntiSpoofingBenchmarks/SpeechAntiSpoofingArena?system=wav2vec2-large-antideepfake) [](https://huggingface.co/spaces/SpeechAntiSpoofingBenchmarks/SpeechAntiSpoofingArena?system=wav2vec2-large-antideepfake) [](https://huggingface.co/spaces/SpeechAntiSpoofingBenchmarks/SpeechAntiSpoofingArena?system=wav2vec2-large-antideepfake) [