public-mirrors/Kokoro-FastAPI
1
0
Fork 0
mirror of https://github.com/remsky/Kokoro-FastAPI.git synced 2025-08-05 16:48:53 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Smaller docker image:

Browse source 
CPU image 5.6Gb -> 4.9Gb
GPU image 14.8Gb -> 9.9Gb

Saved 4GBby using the cuda base image instead of cudnn-runtime. Saved 700MB by deleting the UV cache.

Improved the speed of rebuilding the image by merging the RUN commands into one.
This commit is contained in:
faltiska 2025-07-12 10:25:50 +03:00
parent eac7ab4449
commit 4aad01ee36
2 changed files with 37 additions and 67 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

48
docker/cpu/Dockerfile
View file

@ -1,26 +1,17 @@
FROM python:3.10-slim
# Install dependencies and check espeak location
RUN apt-get update && apt-get install -y \
espeak-ng \
espeak-ng-data \
git \
libsndfile1 \
curl \
ffmpeg \
g++ \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/* \
&& mkdir -p /usr/share/espeak-ng-data \
&& ln -s /usr/lib/*/espeak-ng-data/* /usr/share/espeak-ng-data/
# Install UV using the installer script
RUN curl -LsSf https://astral.sh/uv/install.sh | sh && \
# Rust is required to build sudachipy and pyopenjtalk-plus
RUN apt-get update -y && \
apt-get install -y espeak-ng espeak-ng-data git libsndfile1 curl ffmpeg g++ && \
apt-get clean && rm -rf /var/lib/apt/lists/* && \
mkdir -p /usr/share/espeak-ng-data && \
ln -s /usr/lib/*/espeak-ng-data/* /usr/share/espeak-ng-data/ && \
curl -LsSf https://astral.sh/uv/install.sh | sh && \
mv /root/.local/bin/uv /usr/local/bin/ && \
mv /root/.local/bin/uvx /usr/local/bin/
# Create non-root user and set up directories and permissions
RUN useradd -m -u 1000 appuser && \
mv /root/.local/bin/uvx /usr/local/bin/ && \
curl https://sh.rustup.rs -sSf | sh -s -- -y && \
useradd -m -u 1000 appuser && \
mkdir -p /app/api/src/models/v1_0 && \
chown -R appuser:appuser /app
@ -30,14 +21,9 @@ WORKDIR /app
# Copy dependency files
COPY --chown=appuser:appuser pyproject.toml ./pyproject.toml
# Install Rust (required to build sudachipy and pyopenjtalk-plus)
RUN curl https://sh.rustup.rs -sSf | sh -s -- -y
ENV PATH="/home/appuser/.cargo/bin:$PATH"
# Install dependencies
RUN --mount=type=cache,target=/root/.cache/uv \
uv venv --python 3.10 && \
uv sync --extra cpu
# Install dependencies with CPU extras
RUN uv venv --python 3.10 && \
uv sync --extra cpu --no-cache
# Copy project files including models
COPY --chown=appuser:appuser api ./api
@ -46,14 +32,15 @@ COPY --chown=appuser:appuser docker/scripts/ ./
RUN chmod +x ./entrypoint.sh
# Set environment variables
ENV PYTHONUNBUFFERED=1 \
ENV PATH="/home/appuser/.cargo/bin:/app/.venv/bin:$PATH" \
PYTHONUNBUFFERED=1 \
PYTHONPATH=/app:/app/api \
PATH="/app/.venv/bin:$PATH" \
UV_LINK_MODE=copy \
USE_GPU=false \
PHONEMIZER_ESPEAK_PATH=/usr/bin \
PHONEMIZER_ESPEAK_DATA=/usr/share/espeak-ng-data \
ESPEAK_DATA_PATH=/usr/share/espeak-ng-data
ESPEAK_DATA_PATH=/usr/share/espeak-ng-data \
DEVICE="cpu"
ENV DOWNLOAD_MODEL=true
# Download model if enabled
@ -61,6 +48,5 @@ RUN if [ "$DOWNLOAD_MODEL" = "true" ]; then \
python download_model.py --output api/src/models/v1_0; \
fi
ENV DEVICE="cpu"
# Run FastAPI server through entrypoint.sh
CMD ["./entrypoint.sh"]

56
docker/gpu/Dockerfile
View file

@ -1,29 +1,15 @@
FROM --platform=$BUILDPLATFORM nvidia/cuda:12.8.1-cudnn-runtime-ubuntu24.04
# Set non-interactive frontend
ENV DEBIAN_FRONTEND=noninteractive
FROM --platform=$BUILDPLATFORM nvidia/cuda:12.8.1-base-ubuntu24.04
# Install Python and other dependencies
RUN apt-get update && apt-get install -y \
python3.10 \
python3-venv \
espeak-ng \
espeak-ng-data \
git \
libsndfile1 \
curl \
ffmpeg \
g++ \
&& apt-get clean && rm -rf /var/lib/apt/lists/* \
&& mkdir -p /usr/share/espeak-ng-data \
&& ln -s /usr/lib/*/espeak-ng-data/* /usr/share/espeak-ng-data/
# Install UV using the installer script
RUN curl -LsSf https://astral.sh/uv/install.sh | sh && \
RUN apt-get update -y && \
apt-get install -y python3.10 python3-venv espeak-ng espeak-ng-data git libsndfile1 curl ffmpeg g++ && \
apt-get clean && rm -rf /var/lib/apt/lists/* && \
mkdir -p /usr/share/espeak-ng-data && \
ln -s /usr/lib/*/espeak-ng-data/* /usr/share/espeak-ng-data/ && \
curl -LsSf https://astral.sh/uv/install.sh | sh && \
mv /root/.local/bin/uv /usr/local/bin/ && \
mv /root/.local/bin/uvx /usr/local/bin/
# Create non-root user and set up directories and permissions
RUN useradd -m -u 1001 appuser && \
mv /root/.local/bin/uvx /usr/local/bin/ && \
useradd -m -u 1001 appuser && \
mkdir -p /app/api/src/models/v1_0 && \
chown -R appuser:appuser /app
@ -33,14 +19,9 @@ WORKDIR /app
# Copy dependency files
COPY --chown=appuser:appuser pyproject.toml ./pyproject.toml
ENV PHONEMIZER_ESPEAK_PATH=/usr/bin \
PHONEMIZER_ESPEAK_DATA=/usr/share/espeak-ng-data \
ESPEAK_DATA_PATH=/usr/share/espeak-ng-data
# Install dependencies with GPU extras (using cache mounts)
RUN --mount=type=cache,target=/root/.cache/uv \
uv venv --python 3.10 && \
uv sync --extra gpu
# Install dependencies with GPU extras
RUN uv venv --python 3.10 && \
uv sync --extra gpu --no-cache
# Copy project files including models
COPY --chown=appuser:appuser api ./api
@ -50,18 +31,21 @@ RUN chmod +x ./entrypoint.sh
# Set all environment variables in one go
ENV PYTHONUNBUFFERED=1 \
ENV PATH="/app/.venv/bin:$PATH" \
PYTHONUNBUFFERED=1 \
PYTHONPATH=/app:/app/api \
PATH="/app/.venv/bin:$PATH" \
UV_LINK_MODE=copy \
USE_GPU=true
USE_GPU=true \
PHONEMIZER_ESPEAK_PATH=/usr/bin \
PHONEMIZER_ESPEAK_DATA=/usr/share/espeak-ng-data \
ESPEAK_DATA_PATH=/usr/share/espeak-ng-data \
DEVICE="gpu"
ENV DOWNLOAD_MODEL=true
# Download model if enabled
RUN if [ "$DOWNLOAD_MODEL" = "true" ]; then \
python download_model.py --output api/src/models/v1_0; \
fi
ENV DEVICE="gpu"
# Run FastAPI server through entrypoint.sh
CMD ["./entrypoint.sh"]