English | MP4 | AVC 1920×1080 | AAC 48KHz 2ch | 2h 04m | 407 MB
Train models using distributed data from a variety of mobile devices to classify images, generate text, and do sentiment analysis
Federated learning is revolutionizing how machine learning models are trained. TensorFlow Federated is the first production-level federated learning platform that makes it easy to build mobile device learning-based applications. In this course, you’ll learn the basics of building Federated Learning models that can be gradually improved by decentralized data that comes from a variety of mobile devices while not violating the privacy of end users.
You’ll start by exploring the nature of problems that TensorFlow Federated helps to solve and you’ll install the necessary software. After that, we’ll jump straight into improving an image classification model using a bunch of samples of decentralized data from specially prepared MNIST dataset. Then you’ll start working with text and apply Federated Learning to text generation using a pre-trained model on Charles Dickens’ texts. Next, you’ll handle a text classification problem with TensorFlow Federated where you’ll use a movie reviews dataset.
By the end of this course, you’ll have the practical skills to prepare both datasets and models for Federated Learning as well as the ability to train and evaluate your own models in TensorFlow Federated.
In this course, you’ll learn how to use Federated Learning in practice using a TensorFlow Federated framework on a variety of popular datasets and Deep Learning models.
- Quickly install all the necessary tools to practice Federated Learning on your own machine
- Apply Federated Learning on a variety of common Deep Learning problems
- Gradually improve your pre-trained models using decentralized data
- Discover what to look for when applying the TensorFlow Federated framework in your own projects
- Train and evaluate your own models in a Federated Learning fashion