Dockerfile

FROM pytorch/pytorch:2.1.2-cuda11.8-cudnn8-devel
ENV DEBIAN_FRONTEND=noninteractive

RUN apt-get update && apt-get install -y \
    build-essential \
    python3 \
    python3-pip \
    python3-venv \
    libssl-dev \
    libffi-dev \
    git \
    wget \
    ca-certificates \
    libgl1-mesa-glx \
    libglib2.0-0 \
    python3-dev \
    g++ \
    && rm -rf /var/lib/apt/lists/*

# Create a symlink for python
RUN ln -s /usr/bin/python3 /usr/bin/python || true

# Create a non-root user
RUN useradd -m -u 1000 user

USER user

# Verify CUDA installation path
RUN find /usr/local -type d -name "cuda*"

ENV HOME=/home/user \
    PATH=/home/user/.local/bin:$PATH \
    PYTHONPATH=$HOME/app \
    PYTHONUNBUFFERED=1 \
    GRADIO_ALLOW_FLAGGING=never \
    GRADIO_NUM_PORTS=1 \
    GRADIO_SERVER_NAME=0.0.0.0 \
    GRADIO_THEME=huggingface \
    GRADIO_SHARE=False \
    SYSTEM=spaces

# Set CUDA_HOME environment variable
ENV CUDA_HOME=/usr/local/cuda-11.8
ENV TORCH_CUDA_ARCH_LIST="6.0;6.1;7.0;7.5;8.0;8.6+PTX;8.9;9.0"
ENV LD_LIBRARY_PATH=${CUDA_HOME}/lib64:${LD_LIBRARY_PATH}

# Set the environment variable to specify the GPU device
ENV CUDA_DEVICE_ORDER=PCI_BUS_ID
ENV CUDA_VISIBLE_DEVICES=0

# Clone the RB-Modulation repository
RUN git clone https://github.com/google/RB-Modulation.git $HOME/app

# Ensure CSD directory exists and is in the correct location
RUN if [ ! -d "$HOME/app/third_party/CSD" ]; then \
        echo "CSD directory not found in the expected location" && \
        exit 1; \
    fi

# Set the working directory
WORKDIR $HOME/app

# Upgrade build tools
RUN python3 -m pip install --upgrade pip setuptools wheel

# Download pretrained models
RUN cd third_party/StableCascade/models && \
    bash download_models.sh essential big-big bfloat16

# Install StableCascade requirements
RUN cd third_party/StableCascade && \
    pip install --no-cache-dir -r requirements.txt && \
    pip install --no-cache-dir jupyter notebook opencv-python matplotlib ftfy

# Install gdown for Google Drive downloads
RUN pip install --no-cache-dir gdown

# Download pre-trained CSD weights
RUN gdown https://drive.google.com/uc?id=1FX0xs8p-C7Ob-h5Y4cUhTeOepHzXv_46 -O $HOME/app/third_party/CSD/checkpoint.pth

# Verify the download
RUN if [ ! -f "$HOME/app/third_party/CSD/checkpoint.pth" ]; then \
        echo "CSD checkpoint file not found" && exit 1; \
    fi

RUN ls -la $HOME/app/third_party/CSD

# Ensure CSD is a proper Python package
RUN touch $HOME/app/third_party/CSD/__init__.py

# Update PYTHONPATH
ENV PYTHONPATH=$HOME/app:$HOME/app/third_party:$PYTHONPATH

# Print Python path
RUN python -c "import sys; print('\n'.join(sys.path))"

RUN pip install --no-cache-dir pybind11>=2.12 && \
    pip install --no-cache-dir numpy==1.26.4 && \
    pip install --no-cache-dir transformers==4.31.0

# Verify CSD module can be imported (try different methods)
RUN python -c "from third_party.CSD import model; print('CSD model successfully imported')" || \
    python -c "import sys; sys.path.append('/home/user/app/third_party'); from CSD import model; print('CSD model successfully imported')"

# Verify torch is available before installing GroundingDINO
RUN python -c "import torch; print('torch ok:', torch.__version__)"

# Install LangSAM dependencies
RUN pip install --no-cache-dir segment-anything==1.0

# Important fix: disable build isolation for GroundingDINO
RUN pip install --no-cache-dir --no-build-isolation git+https://github.com/IDEA-Research/GroundingDINO.git

# Install lang-segment-anything without dependencies
# to avoid reinstalling GroundingDINO in isolated build mode
RUN git clone https://github.com/luca-medeiros/lang-segment-anything /tmp/lang-segment-anything && \
    cd /tmp/lang-segment-anything && \
    git checkout 9bdf77715fa70fca96452184abdc64a6666e0d46 && \
    pip install --no-deps -e .

# Install Gradio and accelerate
RUN python3 -m pip install --no-cache-dir gradio==6.13.0 accelerate==1.0.1

# Optional: avoids hot-reload warning on Spaces
RUN pip install --no-cache-dir spaces

# Critical fix: keep NumPy on 1.x for torch/torchvision compatibility
RUN pip install --no-cache-dir --force-reinstall "numpy<2"

# Sanity checks
RUN python -c "import numpy; print('numpy version:', numpy.__version__)"
RUN python -c "import torch; import numpy; print('torch ok / numpy ok')"
RUN python -c "import groundingdino; print('groundingdino ok')"
RUN python -c "import lang_sam; print('lang_sam ok')"

# Copy the app.py file from the host to the container
COPY --chown=user:user app.py .

# Command to run the Gradio app
CMD ["python3", "app.py"]