When I run docker run -it new_image, without any reponse about generative_ai_with_langchain

[danrosher]

This worked for me, alter Dockerfile to be

FROM continuumio/miniconda3:latest
RUN apt-get update && apt-get install build-essential python3-dev -y
COPY langchain_ai.yaml /langchain_ai.yaml
COPY notebooks /notebooks
RUN conda env update -n base -f /langchain_ai.yaml && conda clean -afy
EXPOSE 8888
ENTRYPOINT ["conda", "run", "-n", "base", "jupyter", "notebook", "--notebook-dir=/notebooks", "--ip=*", "--port=8888", "--allow-root", "--no-browser", "--NotebookApp.token=''", "--NotebookApp.password=''"]

Then
docker build -t langchain_ai .
and
docker run -d -p 8888:8888 langchain_ai

Note: build-essential is because c++ is required

then goto http://localhost:8888

[benman1]

Hi @scq000, thanks for reporting, @danrosher thanks for coming up with something that works. the notebook directory and the change to the docker run command were key - thanks for that. I've updated the README and the Dockerfile accordingly. Also, your tag name, langchain_ai, is clearly better than my new_image ;)