Skip to content

Latest commit

 

History

History
63 lines (37 loc) · 1.47 KB

install.md

File metadata and controls

63 lines (37 loc) · 1.47 KB
Raw

Installation instructions

The required dependencies work with python3.6 or python3.7. In addition to this it depends upon R packages which are only available for recent versions of R. Some algorithms require tensorflow, ideally with GPU acceleration.

Anaconda (Windows, Linux, ...)

Depends upon the Anaconda Distribution (https://www.anaconda.com)

   # if you do not have python3.6 or python 3.7 create an environment with the required version of python

conda create -n python36env python=3.6  # install and create an python3.6 environment
conda activate python36env              # use source for linux like environments, for example 'source ~/anaconda3/bin/activate python36env'

   # install packages:
   
conda install  r-essentials r-base

conda install pandas scipy pivottablejs rpy2 matplotlib tqdm requests jupyter seaborn
pip install sklearn gpyopt xgboost sets
  • install tensorflow (optional)
conda install tensorflowXXX keras

Linux

It is known to work and tested with Ubuntu vs 18.04, R vs 3.5.1, python3.6 and tensorflow (1.4.1 and higher)

  • install the R packages with the following steps:
R  # opens R
source("cfg/install_packages.r") # run script and answer the questions, this will install the R packages.
  • install the python dependencies with the following steps:
pip3 install -r cfg/requirements.txt
  • install tensorflow

choose your version of tensorflow (CPU/GPU) and install with:


pip3 install tensorflowXXX keras