Skip to content

Latest commit

 

History

History
74 lines (47 loc) · 3.04 KB

installation.md

File metadata and controls

74 lines (47 loc) · 3.04 KB
Raw

Software Installation Instructions

We recommend installing the following software packages to your local computer for your programming lab assignments. (All these are pre-installed in the cloud platform, but we still recommend installing them on your personal computer to be able to conduct these experiments beyond this course.)

Anaconda

(Anaconda ncludes Python, the Jupyter Notebook, and other commonly used packages for scientific computing and data science.)
Installation instructions are here https://docs.anaconda.com/anaconda/install/

Tensorflow using Anaconda

Install

conda create -n tensorflow python=3.5 #or python=2.7 etc.
or
(Use pip or pip3 version 8.1 or higher in order to install TensorFlow. This must be installed when you installed Anaconda)
pip install --upgrade tensorflow
or
pip3 install --upgrade tensorflow

Activate by issuing the following command:

Windows:

C:> activate tensorflow
(tensorflow) C:>

Mac:

$ source activate tensorflow
(tensorflow) $
Your prompt should change

Issue the appropriate command to install TensorFlow inside your conda environment.

To install the CPU-only version of TensorFlow, enter the following command:
(tensorflow) C:> pip install --ignore-installed --upgrade tensorflow

To install the GPU version of TensorFlow, enter the following command (on a single line):
(tensorflow) C:> pip install --ignore-installed --upgrade tensorflow-gpu

Further instructions are here https://www.tensorflow.org/install/

OpenAI gym

We will use openAI gym based environments for assignments. OpenAI gym provides test and simulation environments for many reinforcement learning problems, in particular, games and robotics. It is useful to try your algorithms on different environments available here, even those not used in the assignments.

Documentation: https://gym.openai.com/docs/

Install using the following commands:

git clone https://github.com/openai/gym
cd gym
pip install -e . # minimal install

Installation script

In case you find it hard to to follow the installation guide for packages. Here is the script that you could run to install all dependencies. By following the steps below, you will create an environment called 'rl', the script is tested only under mac. The script is here install.sh.

  1. open your mac terminal
  2. navigate to the directory of the script, run command

chmod u+x install.sh && ./install.sh

  1. type y for all the questions you will be asked during the installation.
  2. if no error pops up, the installation succeeds.

Every time you code the lab, remember to run the following command before coding

source activate rl

To open the jupyter notebook, run

jupyter notebook