Commit History

  1. Remove datasets import, use pyarrow only

    Verified signature
    2709ffc
  2. Use local parquet reads + soundfile decode (no datasets streaming)

    Verified signature
    efa6dec
  3. Fix: use datasets lib for audio loading (resolves HF references)

    Verified signature
    cee51a1
  4. Fix CAM++ path for bucket mount

    Verified signature
    002c004
  5. Add CAM++ weights for builder

    Verified signature
    f4911f4
  6. Add HF Jobs builder script for T2A parquet

    Verified signature
    1efbcb1
  7. initial commit

    Verified signature
    27bf220