Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization(으)로 돌아가기

4.9

별점

52,380개의 평가

•

5,919개의 리뷰

This course will teach you the "magic" of getting deep learning to work well. Rather than the deep learning process being a black box, you will understand what drives performance, and be able to more systematically get good results. You will also learn TensorFlow.
After 3 weeks, you will:
- Understand industry best-practices for building deep learning applications.
- Be able to effectively use the common neural network "tricks", including initialization, L2 and dropout regularization, Batch normalization, gradient checking,
- Be able to implement and apply a variety of optimization algorithms, such as mini-batch gradient descent, Momentum, RMSprop and Adam, and check for their convergence.
- Understand new best-practices for the deep learning era of how to set up train/dev/test sets and analyze bias/variance
- Be able to implement a neural network in TensorFlow.
This is the second course of the Deep Learning Specialization....

Oct 31, 2017

Thank you Andrew!! I know start to use Tensorflow, however, this tool is not well for a research goal. Maybe, pytorch could be considered in the future!! And let us know how to use pytorch in Windows.

Dec 24, 2017

Exceptional Course, the Hyper parameters explanations are excellent every tip and advice provided help me so much to build better models, I also really liked the introduction of Tensor Flow\n\nThanks.

필터링 기준:

교육 기관: Zhiming C

•Apr 18, 2020

A very good organize course! The knowledge is step by step introduced. From Python can one from scratch a learning code establish. And then the course turns into Tensorflow. Only with this method can man have good feeling about how Tensorflow is processed. Very good course, I strongly recommend it!

교육 기관: Benjamín M

•Apr 19, 2020

Concepts very well introduced and explained, with really good explanations about the intuition behind every topic. It's perfect to be able to apply different techniques knowing what they are good for and when to apply them, and at the same time it also shows where to delve deeper if needed/wanted.

교육 기관: Nityesh A

•Oct 08, 2017

Andrew Ng gives a good satisfactory explanation of the techniques covered in this course. He explains when to use the technique, how to use the technique and how one can implement it in Python and then goes on to give an intuition behind it. I think it should work well for newbies (worked for me).

교육 기관: Tejaswini

•May 25, 2020

I've really enjoyed this course. It gives you a great deal of knowledge and I recommend this to anyone who wants to get an intuition of how to optimise, regularise and perform hyper parameter tuning to make your model learn efficiently. The variety of topics and depth offered was good. Thank you.

교육 기관: Rujuta V

•May 18, 2020

This was an extremely informative course which provided an in-depth knowledge of how Hyper-parameters of Neural Network affect the results and methods of tuning those parameters from improving results. The Programming Assignment provides deeper insights of applying the taught methods effectively.

교육 기관: Gary N

•Mar 05, 2020

This course adds to the first with what you need to make models perform well and fast in practice. Each part of the learning process has possible tuning, tweaks, optimizations to improve performance. The material explains why each tweak works, at least at an intuition level. I have learned a lot.

교육 기관: Eddie C

•Feb 18, 2019

My second AI course certificate from Andrew Ng after I left Taiwan AI Labs. Even though it took me more than 2 months to complete because of my kids' winter vacation and Chinese New Year break. I did learn a lot about how to tune and optimize a Deep Learning network. Keep going to the 3rd course.

교육 기관: Shah M D

•Jan 20, 2019

Great Course. This course does explain some optimisation algorithm with quit a good detail. That is a good part of it. Many less courses explain those algorithms at a level of abstraction an undergraduate student needs. Also, it shows the usage of tensorflow, which is used by major practitioners.

교육 기관: Hoang T H

•Oct 26, 2018

I think it's a great course for those who want to learn about technics related in Neural Network and don't want to know the mathmatical underlying too much, or for those who want to get an intuition or a picture about Neural Network. Thanks Dr. Ng and Coursera a lot for giving me a great course.

교육 기관: Muthu R P E

•Feb 03, 2018

Very good course. We learn the basics of Machine learning and Neural Networks in the earlier course. It works fine when we work with the examples given here, but in real world, our basic program does not work. The tuning process is more important for a successful model. Thanks to Prof Andrew Ng.

교육 기관: Martin S

•Jun 17, 2020

Andrew Ng has a great teaching style, the lectures are always easy to follow and to the point. Weekly quizzes and programming exercises are very well done and help to reinforce the topics a lot.

The programming exercises in week 1 and 2 are very low-level and thus not relevant for real projects.

교육 기관: Steven M

•Mar 11, 2018

I felt like this course picked up specific problems and I was guided through them very well. Including theoretical aspects into the program assignments helped me to understand the concepts as I applied them! I also liked the funny comments every now and then. Great highly recommendable course!

교육 기관: Gopikrishnan M

•Dec 07, 2017

This is a beautiful course that builds on the first one, it gives all the intuitions about various hyperparameters and makes us implement all that in python. Then he when we start working with tensorflow, it all makes sense because we actually know what is going on in the functions that we use.

교육 기관: Samuel Y

•Oct 31, 2019

Incredible course. Very comprehensive, and goes over some awesome, industry-relevant optimization algorithms. Clear examples, programming assignments are extremely helpful, etc. Only things to improve would be to increase the difficulty of programming assignments, and focus more on Tensorflow.

교육 기관: Andrei N

•Sep 21, 2019

The content, examples, assignments, and quizzes are thoroughly developed. All the courses of the specialization share the same notation and lead a student from basic concepts to complex ones helping to develop an intuition on each step. The best course on topic of Deep Learning one could find.

교육 기관: Sanket D

•May 25, 2020

The course teaches many SOTA techniques for tuning hyperparameters, various regularization techniques, various optimization algorithms. Howerver, it would have been great to get a hands-on on hyperparameters tuning in real. Rest, the course is amazing and paves a smooth way to deep learning!!

교육 기관: Mohammad S I

•Mar 26, 2020

I am really glad to learn the tuning and optimization techniques. Hopefully, I can implement them whenever I need. Learning a new framework (TensorFlow) and using it to ease up the bigger calculations was the best thing about the course. Hats off to Andrew NG for designing a course like that.

교육 기관: Mahmut K

•Nov 30, 2018

This second course was great in terms of showing improvements. I would have enjoyed a little more rigorous treatment of why improvements work, but then the course could go on and on... I sill think Andrew can spend a little more time on overcoming overfitting. All in all, excellent balance!

교육 기관: Daniel D

•Aug 20, 2018

It's a great course like the others and quite valuable. I am not quite sure how tensorflow fits into optimization, but I was glad to get a good, handholding kind of introduction to tensorflow as in these courses, I had become accustomed to doing things directly using numpy or MATLAB/Octave.

교육 기관: Intan D Y

•Aug 17, 2018

This course helps practitioner or beginner to know how to tune supporting parameters in order to achieve more efficient/accurate NN. In other words, this course helps me figuring how to optimize the NN design, and I think this is recommended for beginners who like to explore Deep Learning/NN

교육 기관: Nachiket R A

•Apr 15, 2018

This course provided a lot of insight in how to improve accuracy by tuning hyper parameters and also introduced multi-class problems and Deep learning programming frameworks! Awesome specialization to have as it aims to create well rounded expertise in Deep Learning and Neural networks area.

교육 기관: Ekamjot S T

•Jul 07, 2020

After the first course, This course was really important for optimizing the Deep Nets and increasing its accuracy to further heights. The discussion forum also helped in clarification of many intriguing doubts. The assignments were suffice in implementing and understanding the fundamentals.

교육 기관: Малышев Я

•Jun 07, 2020

Отличное продолжение первого курса от этой организации. Множество важных моментов отлично описаны, а задачи по программированию помогают закрепить это на практике. Как отдельный курс наверное не стоит наверное смотреть, но если рассматривать всю специализацию целиком - это отличный продукт!

교육 기관: Vaibhav y

•Dec 12, 2019

Coursera is so amazing to provide an opportunity like this to someone who is living in a 3rd world country with almost no opportunities in a high entry barrier field like Data Science. It is inspiring to see what coursers stands for, providing a learning opportunity to everyone, everywhere.

교육 기관: Sinan G

•May 08, 2018

Nice breadth and depth of relevant topics in this course. Andrew Ng is as always very precise about the issues presented and helps build up our knowledge step-by-step in a super structured way. Nice to work with both Python (semi-raw) models and getting a similar introduction to TensorFlow.