Don’t have a Boss, be the Boss! with Ai Hello world, it’s Siraj. And I’m gonna show you how you can build a profitable business on your own using TensorFlow Google’s widely popular machine learning framework A couple years ago, deep learning started to out perform all others machine learning models when given massive amount of data and google saw that it could use these deep neural networks to improve it services, from gmail to photos, to search. They built a framework called TensorFlow out of a need to let researches and developers work together on AI models. BEST ANIME CROSS OVER EVER and have them scale to billions of people Originally it was made for internal use only but them open sourced it so now anyone can use it we can build a really simple TensorFlow model pretty easily. There are lot of examples online for, say, classification for example all of this comes out to just under a hundred lines of code the pipeline generally follows the pattern of preprocess your data set, build the model then train the model in a training loop ultimately saving the trained model file to your local disk as a binary file. But when our model is fully trained, we want to serve it to our users on a server. That way they can access it from the web and that’s where the TensorFlow Serving Library comes in. Serving implements a server that runs ML models and provides remote access to them. We can do common tasks like prediction and classification of any provided data.