# ============ Common ============
numpy
scipy
pydub
pyyaml
onnxruntime>=1.15.0
gradio[mcp]>=4.0.0

# ============ Audiosplitter only ============
# ASR: faster-whisper (CTranslate2)
faster-whisper>=1.0.0
# Clustering for diarization
scikit-learn>=1.0.0

# ============ Voice Extractor only ============
setuptools
# PyTorch CPU-only
--extra-index-url https://download.pytorch.org/whl/cpu
torch==2.5.1+cpu
torchaudio==2.5.1+cpu
# Audio
soundfile==0.13.1
librosa==0.10.2.post1
ffmpeg-python==0.2.0
# ML models
pyannote.audio==3.4.0
openai-whisper==20250625
speechbrain==1.0.2
silero-vad==5.1.2
# WeSpeaker
wespeaker @ git+https://github.com/wenet-e2e/wespeaker.git
# Deps
s3prl==0.4.17
peft>=0.11.0
transformers>=4.40.0
pytorch-lightning==2.4.0
onnx==1.17.0
matplotlib==3.9.3
rich==14.0.0
tqdm==4.67.1