tao Posted July 12, 2018 Share Posted July 12, 2018 Fast.ai - Part 1 - Lesson 1 - Annotated notes Building a world class image classifier with three lines of Python code The first lesson gives an introduction into the why and how of the fast.ai course, and you will learn the basics of Jupyter Notebooks and how to use the fast.ai library to build a world-class image classifier in three lines of Python. You will get a feel for what deep learning is and why it works, as well as possible applications you can build yourself. Introduction These are my annotated notes from the first lesson of the first part of the Fast.ai course. I’m taking the course for a second time, which means I’m re-watching the videos, reading the papers and making sure I am able to reproduce the code. As part of this process I’m writing down more detailed notes to help me better understand the material. Maybe they can be of help to you as well. Lesson takeaways By the end of the lesson you should know/understand the goal of fast.ai; how to use fast.ai; the “practical, top-down approach” of fast.ai; how to run Python code in Jupyter notebooks; how to apply the basic shortcuts in a Jupyter notebook; why we need GPU and where to access them; how to build your own image classifier using the fast.ai library; the basic building blocks of neural network; what deep learning is and why it works; how you can apply it yourself; the basics of how convolutional networks work; what gradient descent, a learning rate and an epoch are; how to set a good learning rate to train your model. Table of Contents TL;DR Introduction Lesson takeaways Table of Contents The goal of fast.ai The practical, top-down approach of fast.ai How to use fast.ai Fast.ai Part 1 course structure How to run Python code in Jupyter notebooks Jupyter Notebook basics Python 3 Jupyter Notebook shortcuts general shortcuts Code shortcuts Notebook configuration Reload extensions matplotlib inline Why we need a GPU and where to access them Options discussed in the lesson video Some more options My personal experience How to build your own classifier using the fast.ai library fast.ai library A labeled dataset Side note: Getting the data - PDL - Python Download Library Explore a data sample Train our first network architecture data learn learn.fit() Try it yourself What deep learning is and why it works Why we classify AlphaGo Fraud detection Artificial Intelligence > Machine Learning > Deep Learning Infinitely flexible function: A neural network All-purpose parameter fitting: gradient Descent Fast and scallable One more thing Putting it all together: examples of deep learning Digging a little deeper: The basics of Convolutional Networks Infinitely flexible function Gradient Descent and learning rate Putting it all together The goal for this lesson Notebooks used in the lesson Interesting links / Links from the lesson If interested, please take the course < here >. Link to comment Share on other sites More sharing options...
This topic is now archived and is closed to further replies.