Dockerfile.stage1

# https://github.com/tiangolo/uvicorn-gunicorn-fastapi-docker
FROM tiangolo/uvicorn-gunicorn-fastapi:python3.8

# Install PyTorch CPU version
# https://pytorch.org/get-started/locally/#linux-pip
RUN pip3 install torch==1.7.0+cpu torchvision==0.8.1+cpu torchaudio==0.7.0 -f https://download.pytorch.org/whl/torch_stable.html
RUN pip3 install sentence_transformers==0.3.8

# Tried to store the nlp model files in the git repository with git lfs, but
# soon hit lfs bandwidth limits, so trying to embed the nlp model files into the
# docker container when building it as an alternative.
#
# From this stage1 docker image, we will build out our app.
#

RUN python -c 'from sentence_transformers import SentenceTransformer; model = SentenceTransformer("some-model-name-here")'