TensorFlow, the open source software library developed by the Google Brain team, is a framework for building deep learning neural networks. It is also considered one of the best ways to build deep learning models by machine learning practitioners across the globe. In deep learning models, which rely on a lot of data and computing resources, TensorFlow is used significantly.
Given its flexible architecture for easy deployment on various platforms such as CPUs, GPUs and TPUs, TensorFlow remains one of the favourite libraries to get into ML. Its huge popularity also means that tech enthusiasts are on a constant lookout to learn more and work more with this library. While there are many tutorials, books, projects, videos, white papers, and other resources available, we bring you these 10 free resources to get started with TensorFlow and get your concepts clear.
The list is in no particular order.
1| Tutorial By TensorFlow (Website):
What better source than the makers themselves! These tutorials offered by TensorFlow on their website are the perfect resources to get hands-on training. The tutorial begins by helping you training your first neural network based on image classification and progresses forward to use tf.keras, a high-level API used to build and train models. It also contains advanced learnings of text classification, regression and other concepts. You can also learn to save, restore, share and recreate your work.
Click here to take a tutorial.
2| TensorFlow White Paper (Paper):
This preliminary whitepaper by Google researchers talks about programming models and basic concepts of TensorFlow. Titled Large-Scale Machine Learning on Heterogeneous Distributed System, the paper begins with a brief introduction to the concept and goes at length to talk about examples of TensorFlow operation types, implementation, its execution in a single device and multiple devices. Along with other important concepts this paper also has a detailed diagrammatic explanation of the concepts.
Click here to read it.
3| Stanford Course On Tensorflow For Deep Learning Research (PPT):
This course by the Stanford university lets you download notes and slides entirely focused on Tensorflow for deep learning research. The entire course is based on TensorFlow which makes it quite convenient for the user to get a thorough basic understanding of TensorFlow. It also has course material on setting up the TensorFlow, basic operations, TensorFlow optimisers, examples of image classification, reinforcement learning, and much more.
Click here to read it.
4| First Contact With TensorFlow: Get Started With Deep Learning Programming By Jordi Torres (EBook):
This book by Jordi Torres, a professor and researcher at UPC and BSC was penned during a Christmas break to share his knowledge of TensorFlow with his students. It includes a practical approach to learn TensorFlow, starting from the basics, to understanding multi-layer neural networks. It covers in detail concepts such as linear regression, clustering and single-layer neural networks, among others. Though it was launched with an intention to equip his students with TensorFlow basics, it has now gone viral as it was of great help to many students and practitioners. Though it is based on the old TensorFlow release (TensorFlow-0.5.0), it is a good read for introduction to the subject.
Click here to read it.
5| Getting Started With TensorFlow By Giancarlo Zaccone (EBook):
This is one of the best resources to help you get started with TensorFlow engine, a robust, user-friendly, and customisable software library of ML code for deep learning and neural networks. It starts by giving an introduction to the fundamentals, followed by details of creating programs using TensorFlow. It would help you in solving the mathematical concepts, ML and deep learning concepts on the go.
Claim your free book here.
6| Learning TensorFlow By Itay Lieder, Tom Hope, Yehezkel S. Resheff (Ebook):
This book provides an end-to-end guide to TensorFlow, helping you to train and build neural networks for computer vision, NLP, speech recognition, general predictive analytics and others. The book emphasises on hands-on and practical approach to TensorFlow fundamentals before diving into deeper concepts. After reading the book you would be able to get a thorough detail of TensorFlow, build deep learning models, scale TF and deploy TF in production setting.
Click here to read.
7| TensorFlow Tutorial By Bharath Ramsundar (Slides):
These lecture slides by B Ramasundar is an excellent introduction to TensorFlow that draws many parallels between NumPy and TensorFlow codes. He has given details such as NumPy to TensorFlow dictionary, linear regression in TF, gradient computation, and other in his descriptive slides. Get acquainted with the basic of TensorFlow with these slides.
Click here to read.
8| Deep Learning With TensorFlow By Cognitive Class (Online Course):
An initiative by IBM, Cognitive Class aims at democratising access to learning data science and cognitive computing. This course by Cognitive Class focuses on this ideology which is free of cost for ML enthusiasts. Slightly on an advanced level, it is suitable for anyone interested in advancing their skills in machine learning, deep learning and TensorFlow. It covers in-depth resources starting from introduction to TensorFlow to CNN, RNN and other areas. This self-paced course can be taken anytime.
Click here to read.
9| TensorFlow: A System For Large-Scale Machine Learning (Paper):
This paper again by Google brain researchers is a good resource to get an understanding and working in TensorFlow. With various use cases and implementation of various models, this paper tries to describe TensorFlow dataflow model in contrast to existing systems. It also explains image classification, language modelling, and others using TensorFlow.
Click here to read.
10| Free Resources On Github:
There are many resources available on Github that explains the working of TensorFlow. These are from beginner to advanced level ML enthusiast who wishes to explore TensorFlow skills. For instance, this course on Github covers TensorFlow basics, regression, classification, clustering and other details. Whereas, another Github course talks details about simple linear model, CNN, C Keras API and others. Here is another instance of TensorFlow resources on Github.
