Daniil Klimov
110d9cde29
- Fix install_mac.sh: use venv + Python 3.12 (3.14 incompatible with ML libs) - Fix run_mac.sh: activate venv, add CPU thread optimization env vars - Fix agent.py: remove f-string from SYSTEM_PROMPT template (NameError on import) - Add missing deps: sounddevice, pydub, imageio-ffmpeg, omegaconf - Optimize for M1: torch.inference_mode, set_num_threads, OMP/MKL tuning - Switch to qwen2.5:3b for faster LLM responses on Mac - Switch Whisper to medium model with auto compute (small+int8 had poor Russian) - Add initial_prompt for better Russian transcription - Add open_app tool for native macOS app launching - Fix TTS: sanitize Latin text to Cyrillic for Silero compatibility - Fix wake word echo: add cooldown after TTS, reset model state, raise threshold - Make "Слушаю" TTS synchronous to avoid mic interference - Fix train Dockerfile: remove tensorflow/onnx2tf (only ONNX needed), fix deps - Fix train.sh: use wget for dataset download, add --shm-size=2g - Add trained hey_cosmo.onnx wake word model - Add TODO section to CLAUDE.md (ChatterBox TTS, Ollama Modelfile ideas) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
91 lines
3.4 KiB
Docker
91 lines
3.4 KiB
Docker
# Dockerfile для обуч<D183><D187>ния wake word модели openWakeWord
|
||
# Python 3.11 + torch (CPU) — без tensorflow (нам нужен только ONNX, не TFLite)
|
||
FROM python:3.11-slim
|
||
|
||
WORKDIR /app
|
||
|
||
# Системные зависимости
|
||
RUN apt-get update && apt-get install -y --no-install-recommends \
|
||
git wget curl ffmpeg libsndfile1 \
|
||
build-essential python3-dev cmake \
|
||
&& rm -rf /var/lib/apt/lists/*
|
||
|
||
# --- Слой 1: PyTorch (самый тяжёлый, кэшируется) ---
|
||
RUN pip install --no-cache-dir \
|
||
torch==2.5.0 \
|
||
torchaudio==2.5.0 \
|
||
--index-url https://download.pytorch.org/whl/cpu
|
||
|
||
# --- Слой 2: ML-зависимости (без tensorflow!) ---
|
||
RUN pip install --no-cache-dir \
|
||
mutagen==1.47.0 \
|
||
torchinfo==1.8.0 \
|
||
torchmetrics==1.2.0 \
|
||
speechbrain==1.0.3 \
|
||
pronouncing==0.2.0 \
|
||
acoustics==0.2.6 \
|
||
pyyaml "scipy<1.15" scikit-learn tqdm
|
||
|
||
# --- Слой 3: Аудио-аугментация ---
|
||
RUN pip install --no-cache-dir \
|
||
audiomentations==0.43.1 \
|
||
torch-audiomentations==0.12.0
|
||
|
||
# --- Слой 4: Датасеты и ONNX ---
|
||
RUN pip install --no-cache-dir \
|
||
"datasets>=2.20.0" \
|
||
"pyarrow>=15.0.0" \
|
||
webrtcvad \
|
||
onnx \
|
||
onnxruntime
|
||
|
||
# --- Сл<D0A1><D0BB>й 5: openWakeWord ---
|
||
RUN git clone https://github.com/dscripka/openWakeWord /openWakeWord
|
||
RUN pip install --no-cache-dir -e /openWakeWord
|
||
|
||
# Ресурсные модели для feature extraction (melspectrogram + embedding)
|
||
RUN mkdir -p /openWakeWord/openwakeword/resources/models && \
|
||
wget -q -O /openWakeWord/openwakeword/resources/models/melspectrogram.onnx \
|
||
"https://github.com/dscripka/openWakeWord/releases/download/v0.5.1/melspectrogram.onnx" && \
|
||
wget -q -O /openWakeWord/openwakeword/resources/models/embedding_model.onnx \
|
||
"https://github.com/dscripka/openWakeWord/releases/download/v0.5.1/embedding_model.onnx"
|
||
|
||
# Патч train.py: убираем зависимость от onnx_tf/tensorflow (нам нужен только ONNX)
|
||
RUN python - <<'PATCH'
|
||
import pathlib
|
||
|
||
train_py = pathlib.Path("/openWakeWord/openwakeword/train.py")
|
||
text = train_py.read_text()
|
||
|
||
# Заменяем всю функцию convert_onnx_to_tflite на заглушку
|
||
old_func = text[text.find("def convert_onnx_to_tflite("):]
|
||
old_func = old_func[:old_func.find("\nif __name__")]
|
||
new_func = '''def convert_onnx_to_tflite(onnx_model_path, output_path):
|
||
"""Skipped — ONNX-only mode, TFLite not needed."""
|
||
return None
|
||
'''
|
||
text = text.replace(old_func, new_func)
|
||
|
||
train_py.write_text(text)
|
||
print("train.py patched: convert_onnx_to_tflite replaced with stub")
|
||
PATCH
|
||
|
||
# --- Слой 6: piper-sample-generator v2.0.0 (совместим с openWakeWord train.py) ---
|
||
RUN git clone --branch v2.0.0 https://github.com/rhasspy/piper-sample-generator /piper-sample-generator
|
||
RUN pip install --no-cache-dir piper-phonemize || true
|
||
RUN pip install --no-cache-dir -e /piper-sample-generator 2>/dev/null || \
|
||
pip install --no-cache-dir piper-tts
|
||
|
||
# TTS модель (.pt checkpoint) для генерации примеров
|
||
RUN mkdir -p /piper-sample-generator/models && \
|
||
wget -q --show-progress \
|
||
-O /piper-sample-generator/models/en_US-libritts_r-medium.pt \
|
||
"https://github.com/rhasspy/piper-sample-generator/releases/download/v2.0.0/en_US-libritts_r-medium.pt"
|
||
|
||
RUN mkdir -p /data /output /samples
|
||
|
||
COPY entrypoint.sh /entrypoint.sh
|
||
RUN chmod +x /entrypoint.sh
|
||
|
||
ENTRYPOINT ["/entrypoint.sh"]
|