More information would be very helpful, but maybe it is useful for you.
It depends on many different factors, but assuming a couple of things:
A docker containing Tensorflow has a name like 'tensorflow', already contains a python script: you can use: docker run tenorflow 'python'
If the script does not exist yet, you can use the docker assembly based on this image using the docker file, something like strings
FROM tensorflow:latest ADD /some/local/path.py /the/path/on/my/docker
Or you can create / run dockers interactively by doing something line by line:
Docker run -ti tensorflow / bin / bash Which will start the tensorflow container instance and start / bin / bash on it. You now have an interactive shell, so you can put the python script in any way and start it.
source share