anujonthemove/datascience-cpu-py3
A lightweight image with commonly used Python libraries for day-to-day Machine Learning tasks.
321
anujonthemove/datascience-cpu-py3 repository overview
Docker for Data Science
This Docker image is for those who use Python and it's libraries/utils for day-to-day Machine Learning/Data Science tasks which only involves running these workflows on a CPU.
Most of what we explore these days is on Jupyter Notebooks therefore, I have made entry point of this container to be a Jupyter Server. Once started, it will generate a Notebook URL on the interactive terminal, one can copy-paste the same on to the browser and start working!
Installed Libraries
Numerical and Scientific computation
numpy
scipy
sympy
Data manipulation and processing
pandas
Interactive coding
jupyter
Machine Learning
keras
tensorflow
scikit-learn
pandas_ml
Data Visualization
matplotlib
seaborn
bokeh
Note: So far, I have added only the go-to libraries to the image, if you use more useful stuff, feel free to reach out to me on the channels mentioned at the end of this Readme, I'd be happy to add them!
Instructions
Section - 1
If you are new to docker and using an operating system other than Linux, then you can go to Official Docker Page for detailed installation instructions. For convenience, I have copied all the instructions required to setup docker on Ubuntu flavor of Linux in this section. Follow along below.
Setting up docker on Ubuntu 16.04 LTS(Xenial Xerus)
Pull docker image
docker image pull anujonthemove/datascience-cpu-py3
Spawning a docker container from the image
docker run -p 8888:8888 -v /<path-to-directory/volume-to-be-mounted>/:/home/ubuntu/mounted_vol/ -it anujonthemove/datascience-cpu-py3
Parameter meaning
- -p port:port forwards port from inside docker
- -v local_volume:docker_volume mounts a local volume/directory to a running a docker container
- -it docker_image runs the container in an interactive environment
Note: Usually, this step is done only once. After this, we just start an existing docker container from the list of containers . Scroll down for more information!
Starting an existing docker container
Once you have made a container out of the docker image, we can just start stop the same container again and again as per our use.
Find out the CONTAINER ID from a list of containers
docker ps -a
Start the container using CONTAINER ID
docker start -i --attach <container-id>
This would lead you to the shell of the same container just as we had run it for the first time. In our case however, it will start our jupyter server with a new URL.
Start Jupyter Server
jupyter-notebook --ip 0.0.0.0 --port 8888 --allow-root
Tag summary
Content type
Image
Digest
Size
641.7 MB
Last updated
almost 7 years ago
docker pull anujonthemove/datascience-cpu-py3