Prior running the docker build command please clone the repository and copy the solution dir to nlp_architect folder.
Place the following Dockerfile inside nlp-architect and run it from nlp-architect dir.
cd nlp-architect/
docker build -t nlp-architect .
FROM ubuntu:18.04
LABEL maintainer="Intel AI Lab NLP [Steve's working solution]"
RUN apt-get update && \
apt-get upgrade -y && \
apt-get install -y bash build-essential ca-certificates python3 python3-pip git vim && \
rm -rf /var/lib/apt/lists && \
apt-get clean
RUN python3 -m pip install --no-cache-dir --upgrade pip setuptools
ENV LC_ALL=C.UTF-8
WORKDIR /workspace
COPY . /workspace/nlp-architect/
WORKDIR /workspace/nlp-architect/
RUN pip install -e . && \
pip install -r nlp_architect/solutions/absa_solution/requirements.txt
RUN python3 -m spacy download en
ENV BOKEH_ALLOW_WS_ORIGIN=[YOUR LOCAL HOST IP (e.g 10.0.0.1)]:5006
WORKDIR /workspace/nlp-architect/nlp_architect/solutions/absa_solution/
CMD ["python3","ui.py"]
EXPOSE 5006
Running it with or without mounted dir (with embeddings and rerank models) gives the following error:
ValueError('Error: Terms file is empty, no terms to re-rank.', )
Hi @stevesolun
As per our conversation, this error is caused by wrong CSV format.
Please open a PR for the other str decoding bug you discovered.
Thanks
Running the following Dockerfile:
Prior running the docker build command please clone the repository and copy the solution dir to nlp_architect folder. Place the following Dockerfile inside nlp-architect and run it from nlp-architect dir.
Running it with or without mounted dir (with embeddings and rerank models) gives the following error:
ValueError('Error: Terms file is empty, no terms to re-rank.', )
It happens on the third iteration.