Machine Learning

Practitioner

MLPractitioner.com

We are on your path to a practice in Machine Learning and Artificial Intelligence

Consider MLP your  resource and information site for news, tools and techniques to maintain your skillset

Tools for Practitioners of Machine Learning and A.I.

Google Colaboratory

Synopsis

  • ‘Colab’ is a very useful, notebook-centric Cloud-IDE for coding and testing data science projects. It connects to you Google Drive which can be mounted to a currently active Jupyter notebook.
  • There is also a runtime option which includes a GPU for those hefty, Deep Learning projects where you want to stay in your seat while waiting for a big network training stage.
  • It can be used when working within a team, sharing datasets and notebooks and collaborating just as with docs
 
Google Colab

Completely Cloud-Based

Sampling of Colab Feature-Set

One of the great things about Colab is that it is totally free. You can drop notebooks in or create them in colab.  Then, you can share these or push them to Github for a wider share. And, having access to a data science equipped,  Jupyter notebook in the cloud is a major plus. If you share a notebook and that person runs it on their own google colab account, everything will run without regard to dependencies or performance.

 

Wow! Access to a Tesla K80 GPU

  • Free GPU assist for deep learning models
  • Works with Tensorflow
  • Special TPU – ‘Tensor Processing Unit’ for TensorFlow models
  • Nicely integrated into the Jupyter environment – Nothing to setup

 

TensorFlow Support

  • Available by default
  • Connect to a TPU (Tenso Processing Unit) or GPU
  • Acceleration starting at an order of magnitude

 

Visualization Tools

  • Matplotlib
  • Seaborn
  • Plotly
  • Bokeh
  • Altair

 

Access to Data from External Sources

  • Google Drive : Several access methods like PyDrive API
  • Google Sheets are accessible
  • GCP : GCS buckets and files
  • Of course URLLIB for files in the cloud
  • Big Query

 

 

 

Welcome To Colab

 

The notebook view

 

Table of contents from H1 markdown cells

 

 

Altair Graphing

Link to your G-Drive

 

Run cell to mount drive

 

Code Snippets

 

Requirements

Installation / Setup

  • Google account (Not GCP)
  • Browser – preferably Chrome on any modern O/S
  • Google Drive
  • Sign-in to Google account
  • Go to Drive and select ‘new’ google colaboratory

 

Examples

 

Video Demo | ‘Brief’

Click ‘+’ to expand

Google Colab Brief Tour

Courtesy of Google

0 Comments

Submit a Comment

Your email address will not be published.