laion/voiceclap-commercial-genuineness
Audio Classification • Updated
Genuineness and vocal-burst-blend predictors, plus SFT3. Both predictor repositories bundle the frozen VoiceCLAP encoder.
Note 0–6 genuineness predictor; frozen VoiceCLAP-commercial encoder plus MLP, bundled for offline inference.
Note 0–10 contextual vocal-reaction blend predictor; same bundled frozen encoder plus MLP.
Note Timed-script SFT3 voice-acting instrument; the final supervised round used 398,282 selected rows.