Hi guys,
First of all, I’d like to thank the brilliant job of @jeremy and @rachel , thank you for creating this impressive and useful course for us!
Why I wrote these notebooks?
Every time I finished a lesson, I wrote down the learning notes, in order to reinforce the knowledge of the lesson, as Jeremy said, the perfect way to learn a lesson, is to teach others to understand the lesson. In this way, you have to figure out a concept and make sure your understanding is correct. On the other hand, the learning note is a perfect advice for those preparing to learn this lesson, tell them where they might be trapped into, or help them to clarfy some confusing concept.
Before sharing the notes, there’s a suggestion for Jeremy and Rachel, I think it maybe a good thing that before starting the lesson, tell the students what’s gonna covered in the following lesson, in other words, providing a brief table of the lesson, which techniques you are going to learn, which concept you don’t have to struggle with for now, because these will be talked in the next lesson! For example, in lesson1, Jeremy actually wants us to have fun with the notebook, just try the ipynb and see what it can do, that’s enough, we don’t need to figure out the details for now. But because I haven’t learn the lesson2 at that time, I thought I have to figure out every details, other wise I cannot understand the following lesson. Therefore, I’ve wasted to much time in lesson 1 to figure out something I couldn’t understand so far, when I study the lesson 2, I found out Jeremy explained everything！
And that’s why I want to share my learning experience to everyone, tell them what you going to learn in the lesson, explain something coufusing in intution, help them getting started.
The learning note consist of four parts:
- Briefing 简述: what you are going to learn;
- Tasks 任务流程: what you have to do in the coding practice;
3: Important concept explaination 重要概念解释: explain something confusing concept with straightforward words. - Core code presenting 核心代码解释: I implemented the code from the course nbs, and rewrote them from scratch, fixed some little bugs, and explained some important code. I implemented almost 90% code of the nbs, because some part of the original nbs is just for explaination.
So sorry for the English speakers, I wrote these notebooks all in Chinese.
I write the notes in my mother tongue – Chinese will be much faster, maybe someday I’ll translate them in English, but python is international, if you ignore the Chinese words and just focusing on the code, perhaps you would know what I’m trying to say (alright that’s bullshit, just use the google translation).
Tables:
- Lesson2 - Cats vs Dogs, using vgg16 model
- Lesson3 - Some tricks to avoiding over-fitting and under-fitting
- Lesson4 - Part one - More details about CNN, build a CNN from scratch in keras
- Lesson4 - Part two - Collaborative filtering and Recommender system
- Lesson5 - NLP tasks, word embeddings
- Lesson6 - RNNs and LSTM, simple RNN in Theano
- Lesson7 - Nuclear weapon in image classification (ResNet, InceptionV3, and FCN)
Lesson7 is most exciting lesson for me, there’re something coolest things I’ve ever seen, I can’t wait to learn the part 2!
If you have some suggestions for my notes, please let me know, I’d love to hear, and hope you guys enjoy it!
Special thanks:
Thanks @rqyang for your suggestions, you helped me a lot.Thanks @yousheng for providing me the platform to publish my notes.
