Original English: Teach Yourself Deep Learning with TensorFlow
Author: Vincent Vanhoucke,google chief Scientist. Translation: Guokai Han.
In recent years, deep learning has become one of the hottest topics in the field of machine learning. TensorFlow, as an open source project, is our recently released in-depth learning platform. Our goal is to give everyone the ability to use deep learning techniques. We were excited to have more than 4,000 users fork it on Github in a few weeks, and fans from all over the world gave 16000+ collections.
In order to help more engineers and data scientists use the deep learning technology, we and udacity to open a new in-depth study course. This short condensed course provides you with all the basic tools and vocabulary to get started in depth learning and guides you through using it to solve some of the most common machine learning problems. There are also interactive TensorFlow notes for directly reflecting and implementing the concepts presented in the lectures.
The course, which consists of 4 parts, provides a major build guide for solving problems from image recognition to text analysis. The first lesson focuses on basic concepts, understanding the well-established practices of machine learning: Building Your data and experiment programs, and training simple classification models. On these foundations, the second lesson explores how to make these simple models more in-depth and powerful, and to understand the ensuing scalability problems, especially regularization (regularization) and hyper-parametric (hyperparameter) adjustments. The third lesson is the content of convolution network (convolutional networks) and image recognition. The last Lesson IV uses embedded (embeddings) and recurrent neural networks (recurrent neural networks) to explore generic text and sequence models. At the end of the course, you will implement and train the variants of these models on your machine, and you can use this knowledge to prepare yourself for solving your own problems.
Our overall goal in designing this course is to provide a fast, direct way for machine learning enthusiasts to use in-depth learning techniques to solve practical and interesting problems. Now we are very excited to share with you the course we have done. It is interesting to gather these top experts together in Udacity to create and design this online course. For more details please see udacity's blog Register this course . I hope you like it.