Index/ Python

---

Setting Up a Python Machine Learning Development Environment

There are few options

• Option 1: Cloud platforms
• Option 2: Native setup of your machine
• Option 3: Running a docker container
• Finally, get some data. See our data page

Option 1: Cloud Platforms

There are popular hosted ML environments.
This is the quickest way to get going with Python and machine learning.

Supported operating systems : All

1A: Google Colab

We recommend Google Colaboratory.

Pros:

• Nothing to setup on your laptop
• Just need a Google account
• Supports majority of ML libraries
• Provides CPU / GPU / TPU runtime environments
• And important of all It is FREE

Cons:

• Works great with single notebooks, but including other python modules can be tricky

---

Option 2: Native Machine Setup

Here we are going to install and setup the computer.

Supported operating systems : Windows, Mac, Linux

Step 2A: Install Anaconda Python

We highly recommend installing Anaconda Python distribution.
It has pretty much all the packages needed for ML.

Get it from here for your platform

Step 2B: Install Additional Packages

The following extra packages are recommended

• jupyterlab

Install as follows using CLI

    $   conda install jupyterlab

Also can be done via Anaconda console.

Step 2C: Start Jupyter or Jupyterlab

Start as follows

    $   jupyter lab

In browser, go to Jupter page; Usually at localhost:8888

Step 2D: Test the Setup

Use this notebook to see if your environment is setup Testing123 notebook

---

Option 3: Docker Images

There are variety of Docker images that have fully setup ML environments.
Follow the instructions for each Docker

• ElephantScale training image

---

Checklist

At this point, you should

• have access to a Python development environment

---

Index/ Python