Datasets:

---

SpeechAntiSpoofingBenchmarks

/

arena-manifest

like 0

Follow

Speech Antispoofing Benchmarks by lab260 8

Adds SpeechAntiSpoofingBenchmarks/LibriSeVoc (test split) to core_set, pinned at 13d69268a57afc7ebba2421d2a31730a390f1ff2.

• 18,487 trials: 2,641 bonafide (LibriTTS gt) / 15,846 spoof (6 neural vocoders: DiffWave, MelGAN, Parallel WaveGAN, WaveGrad, WaveNet, WaveRNN).
• License CC BY-SA 4.0. Paper arXiv:2304.13085.
• Core addition re-computes coverage for every existing submission.
• Re-ingest after merge to subscribe the dataset for webhook routing.