Intelligent Mobile Projects with TensorFlow

Create Deep Learning and Reinforcement Learning apps with TensorFlow on iOS, Android, and Raspberry Pi About This Book * Build TensorFlow-powered AI apps on smart phones and Raspberry Pi, making AI available anytime anywhere * Covers all major tasks of modern AI: computer vision, natural language and speech processing, and deep reinforcement learning, as well as traditional machine learning * Full of hard-earned practical tips, insights, and working code not available in the TensorFlow documentation and example code that'll help you develop your own mobile AI apps more efficiently Who This Book Is For If you are an iOS and/or Android developer who enjoyed the exciting mobile-first world in the past decade and want to enter the great new AI-first world, or if you are an AI or machine learning engineer or researcher, preferably already familiar with TensorFlow, and want to see how AI can actually run on your mobile devices, anytime and anywhere, or even if you're just new to both mobile and AI development, as long as you know how to program using C/C++/Java/Python, and are not afraid of learning iOS, Android, and TensorFlow, this book is for you. What You Will Learn * Set up TensorFlow Lite and Mobile for iOS and Android development environments * Retrain TensorFlow's image classification model for accurate recognition * Develop object localization using a YOLO model trained with TensorFlow and Darknet * Train a CNN model to do fast neural style transformation and use the model offline in mobile devices * Implement real-time LSTM model for OCR in your own mobile app * Build your own keyword audio recognition model * Use deep learning to generate natural-sounding speech in mobile app * Use Raspberry Pi to build a simple robot that can see, listen, speak, and learn * Develop deep reinforcement learning, made famous by AlphaGo, powered mobile apps In Detail As a developer you need to always keep an eye on and get ready for what's to be trending in the coming years, while keeping your feet on the ground. So what can be better as a developer to learn about the integration of the best two worlds, the present and the future? Google TensorFlow, the leading open-source framework for building AI models and apps, is used to train all the models deployed and running on mobile devices. This book covers 10 projects of implementations of all major areas of AI on iOS, Android, and Raspberry Pi: computer vision, speech and language processing, and machine learning including traditional, reinforcement, and deep reinforcement. With this book, you'll learn how to use existing TensorFlow models and train your own models and develop intelligent iOS/Android/Raspberry Pi apps running those models. This book will help you realize what amazing AI apps you can build on the smart phones people carry every day, and how to build such apps; more importantly, it will help you be better prepared for the new AI world, while not forgetting the mobile world that's going to linger there for years.