Tensorflow is the most popular deep learning/machine learning framework right now. One of the biggest reasons for the popularity of Tensorflow (and my personal favorite) is the portability. A Machine Learning Engineer can create models using Tensorflow on their local machine then deploy those same models to 100s or 1000s of machines. Another reason for the popularity is because the Tensorflow is primarily used with Python. Developers both old and new having been shifting to Python for the last 10 years, which means there is a huge talent pool out there ready to develop in Tensorflow.
The Google Brain team is primarily responsible for releasing the first iterations of Tensorflow (DistBelief prior to release). In 2015 Google released Tensorflow to the open source community and the development has only continued at scale. Considering the importance and popularity of Tensorflow I thought it was a good idea to create a resource list for Tensorflow learning/training/research.
Tensorflow Resources
Research Topics on Tensorflow
Tensorflow – Official site for all things Tensorflow including downloading and installing. Read through the documentation and getting started guide. For a 15 hour deep dive into Tensorflow go through the Machine Learning Crash Course. 15 hours sounds like a lot but break it up into 30 minutes a day for 30 days. After 30 days you’ll have more of an understanding of ML/DL with Tensorflow than most of the competition.
Tensorflow Source Code – At some point in your Tensorflow journey you may want to jump directly into the source code. Tensorflow is an open source project and like most popular open source projects it’s on GitHub.
Hands On Tensorflow Resources
Tensorflow Playground – Interactive Neural Network inside the browser. It allows you to train data from 4 different data sets. You can control features, neurons, learning rate, activation, regularization, etc. One of the easiest things to try is running the same data type through the different activations to see which is faster.
JavaScript Tensorflow? – At first glance I didn’t realize the potential of having a JavaScript Library for Tensorflow. What benefit would come from training models in the browsers? After playing around with some of the demos (Pac-Man) on Tensorflow.js I started to understand how this can open doors to better game develop, human-computer interaction, and more.
Hands-On Machine Learning with Scikit-Learn & Tensorflow – Shamelessly stole this recommendation from a colleague. Should this be on the list for the Big Data Beard Book Club? I think so!
Docker Tensorflow – Super simple way to get started using Tensorflow. Data Engineers can pull Docker tensorflow/tensorflow then pick CPU or GPU to get started developing with Tensorflow. I’ll say it again….a super simple way to get up and coding with Tensorflow. Go download it right now!!
Tensorflow Resources Video
Why Tensorflow is Awesome for Machine Learning – Since I created this list I’m definitely going to put my video at the top of the Tensorflow video. In this video I breakdown Tensorflow was a monumental tool for Deep Learning and Machine Learning.
Siraj Raval YouTube – Siraj Raval has a huge following on his YouTube Channel which is all about Machine Learning, Artificial Intelligence, and Deep Learning concepts. Checkout his first video on Tensorflow in 5 minutes for a quick high level overview of Tensorflow. Then watch my favorite Tensorflow video of creating an image classifier for training a model to detect is this picture of Darth Vader or not.
What is missing? Do you have a suggestion for a resource that should be added? Make sure to put those suggestions for Tensorflow resources in the comment section below.
WANT MORE BIG DATA BEARD?
Find out more tips about Tensorflow, Data Analytics Architecture, and other cool topics in Big Data by subscribing to the Big Data Beard Podcast.
[…] by the advances in the open-source data science world with frameworks like Caffe and Tensorflow. In a previous post we talked about the importance that frameworks like these were built […]