ai-toolkit/docker/Dockerfile

FROM nvidia/cuda:12.6.3-devel-ubuntu22.04

LABEL authors="jaret"

# Set noninteractive to avoid timezone prompts
ENV DEBIAN_FRONTEND=noninteractive

# Install dependencies
RUN apt-get update && apt-get install --no-install-recommends -y \
    git \
    curl \
    build-essential \
    cmake \
    wget \
    python3.10 \
    python3-pip \
    python3-dev \
    python3-setuptools \
    python3-wheel \
    python3-venv \
    ffmpeg \
    tmux \
    htop \
    nvtop \
    python3-opencv \
    openssh-client \
    openssh-server \
    openssl \
    rsync \
    unzip \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/*

# Install nodejs
WORKDIR /tmp
RUN curl -sL https://deb.nodesource.com/setup_23.x -o nodesource_setup.sh && \
    bash nodesource_setup.sh && \
    apt-get update && \
    apt-get install -y nodejs && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

WORKDIR /app

# Set aliases for python and pip
RUN ln -s /usr/bin/python3 /usr/bin/python

# install pytorch before cache bust to avoid redownloading pytorch
RUN pip install --no-cache-dir torch==2.6.0 torchvision==0.21.0 --index-url https://download.pytorch.org/whl/cu126

# Fix cache busting by moving CACHEBUST to right before git clone
ARG CACHEBUST=1234
RUN echo "Cache bust: ${CACHEBUST}" && \
    git clone https://github.com/ostris/ai-toolkit.git && \
    cd ai-toolkit

WORKDIR /app/ai-toolkit

# Install Python dependencies
RUN pip install --no-cache-dir -r requirements.txt && \
    pip install flash-attn --no-build-isolation  --no-cache-dir

# Build UI
WORKDIR /app/ai-toolkit/ui
RUN npm install && \
    npm run build && \
    npm run update_db

# Expose port (assuming the application runs on port 3000)
EXPOSE 8675

WORKDIR /

COPY docker/start.sh /start.sh
RUN chmod +x /start.sh

CMD ["/start.sh"]