What is TensorFlow?

on my bucket list for this year is a commitment to do more tinkering with machine learning and Ai and as I've started this research I keep running into references to what is a tensor what is a tensor flow let's go ahead and break these down real quick tlur is an open source project you can access via GitHub and it can be used with a variety of modern languages like python JavaScript Java and C++ put simply it's used to develop and improve Ai and machine learning models which is why you hear abou

t it so much so the second question makes me ask what is a tensor anyway it turns out a tensor is an algorithm expressed as a multi-linear which is essentially a complex array used in machine learning one way of showing this is a multi-dimensional array kind of like this shape tensil flow is an open Source platform for creating and using bottles is just that simple and what I wanted to outline is kind of the way you work through it it is a very interative platform where you can really get starte

d immediately taking advantage of machine learning it comes first by doing the training now when it comes to building these models creating them there are many decisions to make and this is where tensor flow really AIDS you in this process and you can decide it to say hey I wanted to train this I have a set of data uh I want to make this model be a CPU be trained by CPU or GPU or there's there's even tensor uh Computing processing units that you can do as well to train your data but it gives you

an opportunity to make that decision when you want to create the model now if you don't have your own data you can it will give you data sets that you can leverage all right if you uh kind of don't want to go the process of creating your models by hand there are some starter let's say we call them estimators which are like starter uh neural networks to get started or models that you can use to really try to get jump start your your your process of doing this and then the last part is once you h

ave something you have to deploy that well another great part of that is you could also give you an API test flow is going to give you an API you can use code name caras uh but that be a way that you can actually input data to train your models it encapsulates any other the models that you want to create so very important there that process of getting started is very easy now the last part you may say is all right I'm using tensor flow I want to know what I want to achieve in my application but

how do I get it into where I want to develop and that's where the last part that it offers for you it gives you a method for deployment a it's going to let you get into to mobile and embedded abbreviated em so iOS Android sdks and the ability to get on Raspberry pies all right Len arm sdks for you to get there as well secondly you'll be able to get on browsers there's going to be a library which is tensorflow.js which will give you a mechanism for creating and executing your models inside um you

r code as well and also you can do I'm kind of mention embedded systems but that's something you can do but there's an additional piece here too you can get even get in on embedded devices they even have microcontrollers so the opportunities are endless here to really get started here well for me when I want to get started with something I actually love being able to take advantage of any platform I can that can help me jump start and get and get going and that's what you should do I want you to

leave in the comments what you think about tensor flow and tell us how easy it was to get jump start your journey as always senior ESP spinia