"""
Anvaya · Multi-Expert Speech Pathology & AI Diagnostics Pipeline (ml package)

Research-grade, multi-modal speech diagnosis:
  - ml.data        real-corpus acquisition, unified dataset, speaker-held-out split
  - ml.model       stutter classifier (wav2vec2 + LoRA), pronunciation/articulation
                   (Praat + wav2vec2-CTC GOP), fusion + self-calibration, evaluation
  - ml.cli         single entry point (download / build-dataset / train / eval /
                   diagnose / self-check)

See README.md and docs/ for the reproducibility trail. Intentional, interpretive
decision: `ml/__init__.py` stays a bare module so importing the package never
pulls in torch/parselmouth at import time; consumers import the submodules they
need.
"""
__version__ = "0.1.0"