WHY
The reason why I decided to write this article was that I learned the following books during the year-end and New Year holidays, but there are some parts that get stuck if I am not familiar with Python, so I was able to cut down on mental costs to build an environment and read this book. I was wondering if some people couldn't complete it.
Deep Learning from scratch-The theory and implementation of deep learning learned from Python
WHAT
The part that I thought was clogged was the visualization of the figure. I use matplotlib
to visualize the figure, but there is a trap when building the environment, and if it fits in it, it may take time to solve it.
That's where Docker comes in, where anyone can build the same environment. I have built an environment for docker in Vagrant and use docker, so I have realized an environment completely independent of my personal computer.
The technology used to achieve this is:
These technologies make it easy to carry and understand.
HOW
How to actually do it.
Get the code and move to the code folder
git clone https://github.com/SnowMasaya/deep-learning-excerise.git
cd {Downloaded folder}
Launch Vagrant If you are already using Vagrant, make sure your IP address is unique.
Modify the following part of the Vagrantfile.
config.vm.network "private_network", ip: "{IP address to modify}"
vagrant up
Environment
ansible-playbook -i provisioning/hosts provisioning/site.yml
Get Docker image
docker pull masayaresearch/deep-learning-scratch-excerise
Start Docker and access Docker
docker run -p 8888:8888 -it masayaresearch/deep-learning-scratch-excerise bash
Launch iPython Notebook
ipython notebook --no-browser --port 8888 --ip=*
Access browser
http://192.168.33.25:8888
When the screen below appears, you can see that you can access the ipython notebook.
Please refer to deep-learning-1.ipynb because it actually draws the image.
github
https://github.com/SnowMasaya/deep-learning-excerise/blob/master/docker/python/deep-learning-1.ipynb
Docker hub
masayaresearch/deep-learning-scratch-excerise
https://github.com/oreilly-japan/deep-learning-from-scratch
Recommended Posts