Chevron Left
Retour à Convolutional Neural Networks

Avis et commentaires pour d'étudiants pour Convolutional Neural Networks par deeplearning.ai

4.9
étoiles
37,078 évaluations
4,822 avis

À propos du cours

This course will teach you how to build convolutional neural networks and apply it to image data. Thanks to deep learning, computer vision is working far better than just two years ago, and this is enabling numerous exciting applications ranging from safe autonomous driving, to accurate face recognition, to automatic reading of radiology images. You will: - Understand how to build a convolutional neural network, including recent variations such as residual networks. - Know how to apply convolutional networks to visual detection and recognition tasks. - Know to use neural style transfer to generate art. - Be able to apply these algorithms to a variety of image, video, and other 2D or 3D data. This is the fourth course of the Deep Learning Specialization....

Meilleurs avis

RK
1 sept. 2019

This is very intensive and wonderful course on CNN. No other course in the MOOC world can be compared to this course's capability of simplifying complex concepts and visualizing them to get intuition.

RS
11 déc. 2019

Great Course Overall\n\nOne thing is that some videos are not edited properly so Andrew repeats the same thing, again and again, other than that great and simple explanation of such complicated tasks.

Filtrer par :

126 - 150 sur 4,773 Avis pour Convolutional Neural Networks

par JP L

22 nov. 2017

Extremely well done. Great balance between hand holding/help from the forums and effort in learning. I certainly appreciate the fact that after the course, you are ready to run in the real world working on AI endeavors. They also use all the most recent and up-to-date tools en development environments like Python notebooks, Keras and Tensorflow which makes you immediately proficient working in AI projects. Kudos to the team !

par Souvik S B

20 nov. 2017

This is an excellent course and so far gives best understanding of convoluitonal Network and how it works. But the grading issues needs to be resolved. One thing I specially like about andrew NG courses is how it explains the basics and how algorithms are written from scratch for better understanding. Would be good if we could do the same for YOLO and Facenet.However the assignments are well designed for good understanding.

par Maxime

22 sept. 2020

The course is very interesting but we will have to practice after all that and go through the github codes in detail!

I found the professor Andrew is very clear in his explanations, especially in his desire to visualize what there is behind this complex models.

On the other hand I found the part on the Yolo model a little less well explained especially with regard to the anchor boxes. But I'm going to dig deeper into this.

par michael z

19 sept. 2019

Probably the best course in the specialization and the best course online on ConvNets!

Very engaging and interesting assignments, which cover advanced topics in an approachable manner. teaches current technologies (Keras, TensorFlow). The course goes into some of the math but doesn't get bogged down in it. The course includes recent developments in ConvNets such as the YOLO algorithm, Neural style transfer, and FaceNet.

par Vipul S

9 avr. 2018

Hey,

There are lot of things are happening in computer vision field and this course helped me in understanding the concept like convolution and their use in computer vision field. Practical advice like using existing open-source implementation or existing network architecture are really helpful.

Overall this course equipped me to understand the CNN and it's practical application in computer vision field.

Thanks

Vipul Shaily

par Praphul S

26 nov. 2019

Some exercises very interesting, especially the last week. Why transpose was required made me reflect on the first course's content that dimensions matching will be a very useful technique to debug. Some highlights were the need for the convolution and how it reduces the complexity. The pace of the videos was good and details were very well explained (along with references which encourages to explore more on interest).

par Tao Z

30 mai 2019

Andrew and his teaching assistants made difficult course easy to understand. This is not trivial at all. The exams not only tested students' knowledge but also provide hands on experience on real models, which should be very handy when students want to implement their own AI solutions by themselves later on. Andrew is certainly an excellent teacher and an outstanding AI ambassador, besides being a pioneer in the field!

par Kévin S

31 juil. 2018

You will go deep into image recognition and image processing related to deep learning. As this course show how to use pre-trained model, I should expect to get a model-hub (like docker-hub) like somewhere... but no.

Also I'm not sure to be able to do the exercice outside the notebook, because there is a lot of 'import' and libs to make work. An 'annexe'/'optional' course on how to setup environnement could be nice.

par AVEEK G

22 juin 2020

Superb course structure, the assignments beautifully complement the lectures and the amount of guidance makes it easy even for someone not too acquainted with programming. As a suggestion would have liked slightly organized detailed presentations which would help in reviewing the course material later by glancing through rather than going through the lectures. Over all an awesome course with great learning. Thanks

par Yuwen W

1 avr. 2020

De-mystified sophisticated topics as always. Thru this course, I get a good understanding of the concept and basic building blocks of CNN, and the idea behind object localization, face recognition, neural style transfer.

After this course, I feel there is still a big gap between understanding the concepts and using them in the real world. Will move on to the tensorflow specialization to get more hands-on practice.

par Mohd Z C A

18 janv. 2020

The lectures, quizzes and assignments are designed to help you to understand the topics, not to penalize you. Real-life applications really help me to understand the concepts and the underlying principles. Only one minor issue that I think needs to be addressed - the use of older version of TensorFlow. The latest TensorFlow is not backward compatible and causes major issue when I tried to run the codes locally.

par ANTHONY R

11 nov. 2019

Excellent course with sufficient detail to become instantaneously productive, but at same time more deeper appreciation of internals that must be mastered when beginning designs don't work. Good launch point for learning new DNNs that are part of open source. Much better than Tensor Flow courses that just want you to know how to use the tool. I am ready to tackle my application which is wireless communications.

par Leigh L

14 déc. 2018

This course is a wonderful journey for me. I can certainly apply CNN skills into some of very interesting fields. I have already begun to experience other styles to argument my son's photo. It is a great fun. The facial recognition technique is great to learn. I'm living in China now. Chinese government applies the FR into many public CCTV. It is interesting to observe how they are using it (to say the least :)

par Melvin M

2 sept. 2019

An incredible course about "Convolutional Neural Networks" and related applications to image data. A complete and in-depth course concerning the most important concepts and algorithms about Computer Vision. Furthermore, a fun implementation section which enables youto to create exciting applications ranging from safe autonomous driving, to accurate face recognition, to automatic reading of radiology images.

par Akshay N

22 oct. 2018

Very well structured and informative course. Got to learn plenty of new things, as well as an intuitive understanding of ubiquitous applications like face recognition. The only downside is that for learners not having a hold of frameworks like Tensorflow, the assignments can be a little challenging to tackle. Nonetheless, it helped me glean a very comprehensive understanding of CNNs. Keep up the good work.

par Pui L H (

2 mai 2018

This is a great series of courses. He made things really clear and easy to understand. The assignments examples are so clear and neat. I actually used many assignments as a building block of my machine learning projects in production. I really hope that Dr Andrew Ng will give another series of courses about machine learning again, especially in the reinforcement learning area and the latest technology.

par Qiongxue S

4 mars 2019

I learned a lot from this CNN course, notations, algorithms, tensorflow and keras application. I would strongly recommand to learn this course. It made me think a lot smart applications in daily life and know better about what artifical intelligence is. Of course this is far more than enough, and I will keep learning the related knowledge and reading more about NN. Thanks a lot for the excellent tutorial!

par Rohit K

6 juil. 2019

Hello Andrew, I am a big fan of you. Learning from your every course. Very unfortunate that I can do that remotely only.

One thing that I want to mention - Can we have lecture notes on coursera, just like the way used to in CS229 that we can read before coming to next lecture. I found that that was very useful in understanding when things get harder.

Thanks hope we can improve coursera in that matter.

par Kocić O

15 mars 2018

This course is almost perfect. It gives all the intuition that one might need about ConvNets and it introduces you to the most exciting papers in the field gently and in a fun way. However, in my personal opinion backpropagation of ConvNets should be treated in more details even if that requires some mathematical rigor. One more argument to this is that it can always be made an optional video/assignment.

par Atul A

12 déc. 2017

Excellent course! One of the best courses on ConvNet; it is rigorous and yet fun because of the broad range of projects - from Object Detection to Face Recognition / Face Verification and Neural Style Transfer. Andrew Ng's hallmark is his rigorous and thorough instructions from first principles. I would highly recommend this course to anyone looking to dive deeper into deep learning and computer vision!

par ANGIRA S

31 mars 2018

This can be like the journey where you start as an acquaintance to the CNN's and end as an intimate friend. The excellent thing about this particular course is that it'll introduce you to the seminal computer vision papers and Prof. Ng will also guide as to the difficulty level of the papers. Another amazing learning opportunity is the case study. The text is already online, but the learning is here!

par Vitalija S

30 juin 2020

Loved it but just as others have noted, programming exercises could have more comments about what we are doing because I had to spend lots of time trying to figure out what the task wants me to do. In addition, many links provided in comments about tensorflow documentation don't work. But as I said, this course was amazing because it helped me to understand many important things about CNN. Thank you.

par Rahul M

14 févr. 2018

This is just exceptional. Making cutting edge research accessible to learners. Making tough concepts available and understandable to beginner/intermediate students is hard enough, but Andrew makes it look easy. Some optional assignments where learners do everything from scratch would be good preparation for the real world - maybe this can be part of a capstone added at the end of this specialization.

par Bo M

8 janv. 2018

Some teach so that you understand that they understand. Others teach so that you understand. Andrew Ng belongs to the latter category. The course presents detailed overview of convolutional neural network with concepts ranging from 1D, 2D and 3D convolution, through max and average pooling, to style transfer. All concepts are carefully explained, with great illustrations and easy to follow examples.

par Apperson H J

12 juil. 2020

Course was great (as expected, Andrew is a terrific lecturer) - but it has a couple of problems:

* There are several errors that are pointed out, but sould be fixed in the lecture

* The exercises should use a more recent (ideally current) version of tensorflow

* You need to provide a utility that allows students to download ALL of the material involved (even imagedata that is accessible by links)