NII-SLT26-XLSR-FC: Speech Anti-Spoofing Arena wrapper

Packaged for the Speech Anti-Spoofing Arena from the authors' code and paper.

XLS-R + FC from Ge, Wang and Yamagishi (IEEE SLT 2026): a fully fine-tuned wav2vec 2.0 XLS-R 300M encoder (transformers Wav2Vec2Model), mean pooling of the last hidden state over all frames, and a Linear(1024, 2) head [real, fake]. Trained on CompSpoof V2 (ESDD2) train with the "any component fake = fake" labels, the unmixed "original" class excluded (protocol AnyFakeNoOriginal), seed 2, best-validation-loss checkpoint. Authors' released checkpoint. Input as the authors' collator: 16 kHz waveform, per-clip zero-mean unit-variance normalisation (Wav2Vec2FeatureExtractor), applied after the Arena 4 s cap (first 64600 samples, shorter clips repeat-padded; the authors score whole clips). score = logit[real] - logit[fake] (the authors' Score_real - Score_fake on softmax probabilities is a monotone function of it). FP32 (the authors' eval precision; TF32 off).

  • Weights: re-hosted here under the upstream license (nii_slt26_xlsr_fc.safetensors, commit-pinned); upstream: https://github.com/nii-yamagishilab/AntiDeepfake-AudioLLM/tree/2d445b525bb52518448b353fe1fee51d5f815af9
  • Params: 315.44 M. Trained on CompSpoof V2 train: bona fide CommonVoice, LibriTTS and english-conversation-corpus speech mixed with environmental sound; spoof speech incl. ASVspoof 5, SONAR, CD-ADD, LibriSeVoc and MLAAD sources. The Arena marks ASVspoof5, SONAR, CD-ADD and LibriSeVoc as n/a* (training overlap). Clean studio bona fide speech is out of domain for these checkpoints; several cells are near-random or inverted and are kept (maintainer's choice).
  • Score: higher = bona fide. J-SPAW_LA was scored with the PyTorch wrapper; most other datasets with the TensorRT drop-in class (FP32 on the RTX 4070 Ti SUPER; PyTorch on the RTX 5060 Ti (no engine built there)), which matched PyTorch on the full J-SPAW_LA set (|dEER| <= 0.25, Spearman >= 0.999); the RTX 4060 Ti cards ran PyTorch.

Arena results

arena tier arena rank

EER% (lower is better); EmoSpoofTTS and LRLspoof are spoof-only and use 1-SRR% at the model's DeepVoice EER threshold. Per-dataset scores are under .eval_results/.

EER% 19.1 on ASVspoof2019_LA EER% 29.3 on ASVspoof2021_LA EER% 27.18 on ASVspoof2021_DF EER% 38.56 on ASVspoof5 EER% 73.37 on InTheWild EER% 19.33 on ADD22_eval_31 EER% 29.94 on ADD2023_track12_test_r1 EER% 35.45 on CFAD EER% 34.33 on CD-ADD EER% 38.79 on CVoiceFake_small EER% 31.05 on DeepVoice EER% 34.46 on LibriSeVoc EER% 59.85 on SONAR EER% 40.91 on ArAD EER% 13.27 on DECRO EER% 36.76 on J-SPAW_LA EER% 40.16 on ODSS EER% 42.28 on HABLA EER% 9.54 on DFADD EER% 30.82 on PyAra EER% 35.57 on XMAD EER% 58.54 on EmoFake_test SRR COMPLEMENT 94.84 on EmoSpoofTTS SRR COMPLEMENT 50.72 on LRLspoof

Citation

@inproceedings{ge2026what,
  title     = {What Do Deepfake Audio Detectors Hear? From Self-Supervised Models to Audio LLMs},
  author    = {Ge, Wanying and Wang, Xin and Yamagishi, Junichi},
  booktitle = {IEEE Spoken Language Technology Workshop (SLT)},
  year      = {2026}
}
Downloads last month
180
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Paper for SpeechAntiSpoofingBenchmarks/NII-SLT26-XLSR-FC