I started to feel the limits of doing machine learning on my Windows PC, so I wanted to transfer it to the cloud and use various software and libraries freely with high performance and Linux, so I was able to do it for the time being while researching various things. Instead of a memorandum for the time being.
Since I want to do it on Google Cloud Platform because I use Big Query (BQ), the cloud uses GCE as the title suggests. I've been using Jupyter Notebook since I was doing it on Windows, so I want to use it in the cloud as well. But with Google Cloud Datalab, Python is hard to use in 2.x. .. .. So I decided to put Jupyter Notebook on GCE.
https://console.cloud.google.com/compute/ Go to "VM Instance> Create Instance"
It's very easy to do. The points to note are as follows -The default boot disk is Debian, so change to Ubuntu 16.04 LTS -CPU / memory seems to be adjustable later, so it's appropriate -Allow HTTP / HTTPS traffic because I want to access it with Jupyter Notebook -Set up a firewall as well
After that, start the instance and select "Connect"> "Open in browser window" to make an SSH connection.
For the time being, so that you can use Jupyter Notebook & machine learning. The rest is a library for connecting to BigQuery.
From the command line, just do the following (Port should be set by firewall)
jupyter notebook --ip=0.0.0.0 --port=8888 --no-browser
After that, if you access with ip: port, you will reach the usual Jupyter Notebook screen. It seems that you can also set token and password.
[Set up Anaconda + IPython + Tensorflow + Julia on a Google Compute Engine VM] (https://haroldsoh.com/2016/04/28/set-up-anaconda-ipython-tensorflow-julia-on-a-google-compute-engine-vm/)