FROM nvidia/cuda:11.0-runtime-ubuntu20.04 WORKDIR /home/user RUN apt-get update && apt-get install -y software-properties-common && add-apt-repository ppa:deadsnakes/ppa RUN apt-get update && apt-get install -y python3.7 python3.7-dev python3.7-distutils python3-pip curl wget git pkg-config cmake swig ENV LC_ALL=C.UTF-8 ENV LANG=C.UTF-8 # Set default Python version RUN update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.7 1 RUN update-alternatives --set python3 /usr/bin/python3.7 # Install PDF converter RUN wget --no-check-certificate https://dl.xpdfreader.com/xpdf-tools-linux-4.03.tar.gz && \ tar -xvf xpdf-tools-linux-4.03.tar.gz && cp xpdf-tools-linux-4.03/bin64/pdftotext /usr/local/bin RUN apt-get install libpoppler-cpp-dev pkg-config -y --fix-missing RUN apt-get install tesseract-ocr libtesseract-dev poppler-utils -y # copy code COPY haystack /home/user/haystack # install as a package COPY setup.py requirements.txt README.md /home/user/ RUN pip3 install numpy scipy Cython # Install PyTorch for CUDA 11 RUN pip3 install torch==1.7.1+cu110 -f https://download.pytorch.org/whl/torch_stable.html # Install faiss separately as building latest versions can cause trouble with swig RUN pip3 install faiss-cpu==1.6.3 RUN pip3 install -r requirements.txt RUN pip3 install -e . # download punkt tokenizer to be included in image RUN python3 -c "import nltk;nltk.download('punkt', download_dir='/usr/nltk_data')" # create folder for /file-upload API endpoint with write permissions, this might be adjusted depending on FILE_UPLOAD_PATH RUN mkdir -p /home/user/file-upload RUN chmod 777 /home/user/file-upload # copy saved models COPY README.md models* /home/user/models/ # Copy REST API code COPY rest_api /home/user/rest_api # optional : copy sqlite db if needed for testing #COPY qa.db /home/user/ # optional: copy data directory containing docs for ingestion #COPY data /home/user/data EXPOSE 8000 # cmd for running the API (note: "--preload" is not working with cuda) CMD ["gunicorn", "rest_api.application:app", "-b", "0.0.0.0", "-k", "uvicorn.workers.UvicornWorker", "--workers", "1", "--timeout", "180"]