Chevron Left
返回到 Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization

學生對 deeplearning.ai 提供的 Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization 的評價和反饋

4.9
57,364 個評分
6,590 條評論

課程概述

In the second course of the Deep Learning Specialization, you will open the deep learning black box to understand the processes that drive performance and generate good results systematically. By the end, you will learn the best practices to train and develop test sets and analyze bias/variance for building deep learning applications; be able to use standard neural network techniques such as initialization, L2 and dropout regularization, hyperparameter tuning, batch normalization, and gradient checking; implement and apply a variety of optimization algorithms, such as mini-batch gradient descent, Momentum, RMSprop and Adam, and check for their convergence; and implement a neural network in TensorFlow. The Deep Learning Specialization is our foundational program that will help you understand the capabilities, challenges, and consequences of deep learning and prepare you to participate in the development of leading-edge AI technology. It provides a pathway for you to gain the knowledge and skills to apply machine learning to your work, level up your technical career, and take the definitive step in the world of AI....

熱門審閱

CV
2017年12月23日

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.

XG
2017年10月30日

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.

篩選依據:

5526 - Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization 的 5550 個評論(共 6,515 個)

創建者 Crawford F

2020年12月7日

The final lab is somewhat confusing in that the TensorFlow syntax is poorly explained and the results for the final module would be well served by noting what your first epoch should be as well as the 100th (I spent a long time trying to find non-existant bugs because I had misread the output of my model as epoch 100!!).

Other than that excellent as ever.

創建者 Satyam k

2020年8月18日

This course provide very deep and good knowledge that how to increase speed of your neural network and how we do hyperparameter tunning. But one thing lags in this course is that it won't provide much knowledge about frameworks like Tensorflow and people face difficulty while doing programming exersice because tensorflow knowledge is not provide in depth

創建者 Vishak A

2020年5月14日

I wish more of TensorFlow had been included in the course content. Aside of that major point, I wish the complex mathematical portions had been explained with more precision and codes like "X[0][0]" had been explained with more precision as well. But overall, I think it was hugely worth learning all the thoroughly taught concepts and I am very grateful.

創建者 chinmay h

2020年5月8日

Topics are explained very well. There may be a false sense of accomplishment coming after doing the assignments, which are pretty straightforward. I am going to add in personal tasks which might help me understand the topics more in depth. On a similar front, could you add in a video explaining what to do next. And I don't mean the next course in line.

創建者 jim

2017年11月8日

gain quite a lot of insight into the deep neural network, the tunning, regularization and so on.

one remark on this course, we talked a lot about tunning processes in wk3. However, not much practical exercises on this part, e.g. we didn't try to implement the batch normalization ourselves and to incorporate batch normalization with other parameters etc.

創建者 Aurangazeeb A K

2019年9月30日

Although I loved this course, I believe there are certain parts that could be broken down into even simpler intuitions. If such a change a possible, this course will be the best one out there. Anyway, I really enjoyed the course and it was a great learning experience. Tensorflow was introduced very finely and it aroused my curiousity to learn more.

創建者 Manish.M

2020年3月22日

Really informative course to learn about the various kinds of optimizations and the differences between the optimization techniques. Learnt how to tune the hyper parameters for effective training . Also got a chance to learn about mini-batches and the corresponding gradient descent and the difference between batch and mini-batch gradient descent.

創建者 Alejandro F

2020年2月3日

Un curso muy bueno, el instructor tiene dominio del tema y sobre todo el final del curso es muy bueno en cuestión de poner en practica la teoría que al principio te explica. En ocasiones el instructor va un poco rápido en los términos teóricos y puede llegar a abrumarte. Creo quería ideal poner más ejemplos prácticos cada que explica un concepto.

創建者 Yix L

2019年12月20日

Materials are good and Professor Andrew presents the course in the really understandable level, so I still learn a lot throughout the course even if I have taken similar mooc courses on other platforms. Programming Assignments are much easier than the fourth course (Convolutional NN), but it gives many inspiration to me. Great thanks to the team!

創建者 Hans E

2018年2月18日

Great material, very clear and pleasant teaching, good software environment for the programming exercises. The exercises are a bit boring at times (cut and paste without much thinking) but maybe this is a quick way to memorize the material...

Some long known problems in the exercises should REALLY REALLY be addressed! (would have given 5 stars)

創建者 Guoqin M

2018年6月29日

Content is great! A good introduction to a lot of hyper-parameters in neural net. However, there are some bugs in the evaluation system of programming assignments. For example, the system does not recognize Pythons '-=' operation and gave me a fail, which I did not figure out until I saw the forum where people were having the same trouble.

創建者 Malav A

2020年5月4日

The course was very good. Things were implemented and taught well and at the correct pace. However, while completing the exercise, we can never write the whole code, we have to only edit a few lines of codes. That's not bad for a beginner, but it would have been better if a little understanding about that part of code could be given too.

創建者 Nikola J

2018年5月19日

Andrew is great at teaching. Quality of education is absolutely for 5 stars, but I am giving 4 because of technical difficulties with Jupiter notebook. Often happened that I wrote some code and it could not save, it just displayed error, so I had to copy code to my notepad and rerun the Jupiter notebook, and than copy the code back.

創建者 Ozan G

2020年8月9日

I really like the content but I believe that it is about time the final assignment of this course is updated to Tensorflow 2. There is no point in enforcing learning outdated software... For the massive revenue that this course is generating, the minimal effort to update one Jupyter Notebook should not be too much of a burden...

創建者 Usama B N

2020年5月19日

The course was a very focused approach towards introducing and familiarizing us with the importance of tuning hyperparameters and their impact on the performance. Although, I personally feel like the Tensorflow exercise could have been more detailed and could have used more explanation. I found that exercise somewhat confusing.

創建者 Guoliang

2020年4月3日

The explanation is just as good as the previous course. The reason I give 4 star is that the notebook use TF version 1 instead of 2. Given syntax of 1 and 2 shows great difference, at least I believe so, it would be better that the notebook can be updated. For the rest of the course, very good!!! Suitable for beginners in DL.

創建者 Ytsen d B

2017年8月15日

This course is well taught.

Andrew Ng takes you through the material without error and in a very acceptable pace.

The exercises are very do-able.

They do not challenge hard, but take you by the hand and show you how to implement and improve your neural networks.

The final assignment is a very good tutorial on TensorFlow actually :)

創建者 Emmanuel T

2019年10月3日

Compared to previous module, this one was more of a cookbook and I expected more mathematics in terms of why each optimization work.

Overall, it was still a very interesting hands on approach, finishing with TensorFlow is a bit more difficult to apprehend as all the previous exercices were done in a very different way (Numpy).

創建者 Pawel P

2020年12月9日

Most of the course is great, good overview of different methods and techniques with practical examples. However the TensorFlow programming part is rather confusing, lacking in sufficient explanation of the syntax and overlapping names of python and tensorflow variables which end up producing near impossible to debug errors.

創建者 Girish G

2020年4月13日

This is an amazing course which dwells into the nuances of fine tuning your neural network model. The content of the course is too good. Programming assignments was a bit off. It was really straightforward. Programming assignments could have been more challenging. This will make sure that the concepts are learned properly.

創建者 Le H L

2018年6月10日

The content is generally great and helpful, but the grader did not show me why the result is incorrect, and i constantly had to reload jupyter notebook. I think there should be less template for the exercise so that we have more thinking to do, but the expected result should be maintained so that we know what we did wrong.

創建者 Rakesh S

2017年8月31日

The course explains the reasons and intuition behind tuning hyperparameters and why/how regularization techniques work well when training on large data sets. The only reason I am giving this a 4 star is because the tensorflow introduction seems a little too sparse and could be done better.

Thanks again, team deeplearning.ai

創建者 Juan P A A

2020年1月27日

The contents are actually good, and it doesn't require a very extensive prior knowledge, so it's even suitable for people with little experience in programming or math. However, despite being a course that has been out for over 2 years, there are still some subtitle issues (in English), and typos on a clarification slide.

創建者 John H

2017年8月24日

Well explained..sometimes jumps a bit. I felt lost a couple of times. But I got through it and I'd say this is deifnitely one of the top courses out there.

If they included some optional videos on how this could relate to having a career in this area that'd be very helpful (i.e. what level we need to be able to code at).

創建者 Anmol K

2020年6月16日

This course continues to build on foundations from course 1 of the specialization. Hyperparameter tuning and Regularization methods are quite imperative for optimizing ML models. This course covers these concepts in addition to providing a good foundation for Tensorflow library. Overall, a good course by Prof. Andrew!