In order to learn reinforcement learning, we will proceed in about 100 times. The first 10 times will be the preparation of the environment.
At the beginning, it is better to use Google Colaboratory, so it may be easier to proceed to reinforcement learning 28 or later. Please note that the version of Google Colaboratory may be old. (Added December 4, 2019)
The following is the case of building an environment on a personal computer. Install Python.
The first start was a laptop 10 years ago. ASUS ASPIRE1410 CPU Celeron SU2300 RAM 2GB HDD 250Gb
Very old and low spec. OS Lubuntu 18.04
Despite this low spec, it is largely due to the power of Lubuntu that it moves smoothly.
Python comes with 3.6.8 from the beginning. The command adds 3 like python3.
This is fine, but it also installs Anaconda. Download and install from Anaconda. This area is written on various sites, so I will omit the details. At the end of Anaconda, I was asked if I would like to conda init, so I answered yes.
conda -V
jupyter notebook
Check with. In a web browser http://localhost:8888 Open with. There are many ways to use it.
Jupyter notebook makes development easier.
With the installation of Anaconda, the base is now 3.7.4.
By the way, this machine worked smoothly so far.
Recommended Posts