Commit History

  1. Fix: Apply compatibility patches before speechbrain import

    e53bdb6
  2. Prioritize root in sys.path and add debug info

    35456ff
  3. Fix Import Error: Add __init__.py files and debug logs

    d72a730
  4. Fix ModuleNotFoundError by adding root to sys.path

    08f0d4c
  5. Refactor project structure to Unified AI Voice Detection System

    b3f89f5
  6. Fix SpeechBrain compatibility with newer torchaudio by mocking list_audio_backends

    0b7673f
  7. Fix model loading error by updating torch version and improving logging

    6a16e92
  8. Heavy & Accurate: Integrated SpeechBrain VAD + MMS-300M pipeline

    62f98bb
  9. Upgrade: Switch to MMS-300M (XLS-R) for robust multilingual deepfake detection

    cb14a1d
  10. Fix: Require heuristic evidence for AI verdict on non-English audio to prevent false positives

    0d64f25
  11. Implement Conservative Consensus: SNR Analysis and Human Liveness Veto to eliminate false positives

    f95b090
  12. Implement supreme Hybrid AI detection: sliding window, spectral and temporal features, and language-aware confidence.

    04f3da2
  13. Add comprehensive debug logging to track AI classification confidence calculations

    9d2870b
  14. Upgrade to specialized deepfake detection model and optimize audio processing

    f07936b
  15. Fix: Align prediction label with pitch variance heuristic

    b1df80d
  16. Optimize: Truncate audio to 5s & add per-step timing logs

    28dad85
  17. Fix: Final audio processing fixes and enhanced logging

    a218549
  18. Fix: Add debug logging and robust padding to audio processing

    1f10fcd
  19. Fix: Robust Base64 decoding

    4b4453a
  20. Initial commit: Multilingual Voice Detector with XLS-R Support

    2a4d245