Python Software

The department is providing the free version of Anaconda for Python. Below is the steps for setting up a python environment on a Windows computer.

Setup

Please create the .condarc file when using department PCs, so that conda files are stored in a preferred location.

• [Open] Anaconda Command Prompt
• conda config --add channels conda-forge
• notepad .condarc
  

  channels:
     - conda-forge
     - defaults
  envs_dirs:
     - S:\Documents\conda\envs
  pkgs_dirs:
     - S:\Documents\conda\pkgs 

• {type} exit (Closing the Command Prompt)
• [Open] Anaconda Command Prompt

Option 1

• (In a new Anaconda prompt window)
• conda create -y --no-default-packages -n env-stats
• conda env list
• conda activate env-stats
• conda install -y -c conda-forge tensorflow python=3.7.2 tensorflow-gpu=2.1.0 keras numba (this will take a while ~ 1 hour)

Option 2

• (In a new Anaconda prompt window)
• notepad stats.yml
  name: env-stats
channels:
- conda-forge
- defaults
dependencies:
- numba
- python=3.7.2
- tensorflow-gpu=2.1.0
- tensorflow
- keras
prefix: s:\documents\conda\envs\env-stats
• conda env create -n env-stats -f stats.yml (this will take a while ~ 1 hour)

Usage

• [Open] Anaconda Navigator
• [Click ] (No, don’t show again) to update Anaconda Navigator
• [Change] Applications on [env-stats]
• [Click] [Launch] Spyder
• Run example code at: https://www.geeksforgeeks.org/running-python-script-on-gpu/

Useful Conda Commands

• conda env list — Show all available conda environments
• conda activate ENV-NAME — Selects conda environment ENV-NAME
• conda env export > ENV-NAME.yml — Exports current environment to file ENV-NAME.yml
• conda list — Shows all conda packages installed for active environment
• conda search NAME — Searches for conda packages matching NAME
• conda install -y NAME — Installs conda packages matching NAME
• conda deactivate ENV-NAME — Deselects conda environment ENV-NAME
• conda env remove -n ENV-NAME — Deletes conda environment ENV-NAME