TensorFlow

TensorFlow

The name TensorFlow comes from the operations that such neural networks perform on multidimensional data arrays that are referred to as tensors. TensorFlow is an open source machine learning framework that is used to implement machine learning and deep learning applications faster and easier. It was originally developed by Google. The main benefit to this software is abstraction. Instead of dealing with all the details of implementing algorithms, it focuses on the overall logic of the application and takes care of details behind the scenes.

History

Google first built something called DistBelief during 2011 which is a machine learning system based on deep learning neural networks. Some computer scientist were given the task to make this software faster and more robust and over time this became TensorFlow. The first version (1.0.0) of TensorFlow was released on February 11, 2017. Then another version was released in October 2019. This software runs on CPUs, GPUs, 64-bit Linux, macOs, Windows, and mobile platforms such as Android and iOS. TensorFlow without API is backwards compatible with C++, Go, Java, JavaScript, and Swift with third party packages available for C#, Haskell, Julia, R, Scala, Rust, OCaml, and Crystal. With the API, it is stable with Python and C.

What it can do

TensorFlow can train and run deep neural networks for handwritten digit classification, image recognition, word embeddings, recurrent neural networks, sequence-to-sequence models for machine translation, natural language processing, and PDE (partial differential equation) based simulations.