How to install pytorch in docker file on NVIDIA Jetson Xavier NX

Published

As title mentioned, since I want to form a docker image which include my application using pytorch, the following is my Dockerfile:

From python:3.6-slim

RUN pip3 install numpy==1.19.4
RUN pip3 install opencv-python-headless
RUN pip3 install grpcio
RUN pip3 install protobuf

RUN apt-get update

RUN apt-get install -y libopenblas-base libopenmpi-dev libopenmpi2

#RUN apt-get install -y libgfortran3 libhwloc5 libibverbs1 libhwloc-plugins

#COPY libopenmpi2_2.0.2-2_arm64.deb libopenmpi2_2.0.2-2_arm64.deb

#RUN dpkg -i libopenmpi2_2.0.2-2_arm64.deb

COPY torch-1.8.0-cp36-cp36m-linux_aarch64.whl torch-1.8.0-cp36-cp36m-linux_aarch64.whl

RUN pip3 install torch-1.8.0-cp36-cp36m-linux_aarch64.whl

WORKDIR /app

...

I have already tried any of the above installation, but always got

OSError: libmpi_cxx.so.20: cannot open shared object file: No such file or directory

Is there anyone having any idea on this issue? Please share with me!

Source: Docker Questions

Published
Categorised as docker, nvidia-jetson, pytorch Tagged , ,

Answers

Leave a Reply

Still Have Questions?


Our dedicated development team is here for you!

We can help you find answers to your question for as low as 5$.

Contact Us
faq