Dockerfile.strixhalo

FROM rocm/dev-ubuntu-22.04:latest

# Set environment variables
ENV PYTHONDONTWRITEBYTECODE=1
ENV PYTHONUNBUFFERED=1
ENV DEBIAN_FRONTEND=noninteractive
# Set the Hugging Face home directory for better model caching
ENV HF_HOME=/app/hf_cache

# Install system dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
    build-essential \
    libsndfile1 \
    ffmpeg \
    python3 \
    python3-pip \
    python3-dev \
    python3-venv \
    git \
    rocm-ml-libraries \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

# Create a symlink for python3 to be python for convenience
RUN ln -s /usr/bin/python3 /usr/bin/python

# Set up working directory
WORKDIR /app

# Copy requirements first to leverage Docker cache
COPY requirements-strixhalo-init.txt ./requirements-strixhalo-init.txt
COPY requirements-strixhalo.txt ./requirements-strixhalo.txt

# Upgrade pip and install Python dependencies
RUN python3 -m pip install --no-cache-dir --upgrade pip && \
    pip3 install --no-cache-dir -r requirements-strixhalo.txt && \
    python3 -m pip install --no-cache-dir --force-reinstall -r requirements-strixhalo-init.txt

# Patch s3tokenizer dtype issues for torch 2.9+ compatibility:
# 1. Cast numpy wav to float32 in _prepare_audio (numpy defaults to float64)
#    This fixes the cascade: float64 wav → float64 STFT → float64 mel → assertion failure
# 2. Cast _mel_filters to match magnitudes dtype (guards against float64 checkpoint weights)
RUN sed -i \
    's/wav = torch\.from_numpy(wav)$/wav = torch.from_numpy(wav).float()/' \
    /usr/local/lib/python3.10/dist-packages/chatterbox/models/s3tokenizer/s3tokenizer.py && \
    sed -i \
    's/mel_spec = self\._mel_filters\.to(self\.device) @ magnitudes/mel_spec = self._mel_filters.to(self.device).to(magnitudes.dtype) @ magnitudes/' \
    /usr/local/lib/python3.10/dist-packages/chatterbox/models/s3tokenizer/s3tokenizer.py

# Patch voice_encoder dtype issue: melspectrogram() returns numpy float64 → LSTM requires float32
RUN sed -i \
    's/utt_embeds = self\.inference(mels\.to(self\.device),/utt_embeds = self.inference(mels.to(self.device).float(),/' \
    /usr/local/lib/python3.10/dist-packages/chatterbox/models/voice_encoder/voice_encoder.py

# Copy the rest of the application code
COPY . .

# Create required directories for the application (fixed syntax error)
RUN mkdir -p model_cache reference_audio outputs voices logs hf_cache

# Expose the port the application will run on
EXPOSE 8004

# Command to run the application
CMD ["python3", "server.py"]