A really good course that builds up the knowledge over the concepts covered in Course 1. All the ideas are applicable in real world scenario and this is what makes the course that much more valuable!
great introductory stuff, great way to keep in touch with tensorflow's new tools, and the instructor is absolutely phenomenal. love the enthusiasm and the interactions with andrew are a joy to watch.
автор: Eulier A G M•
The course is marvelous explain and with clear, concise & straight forward concepts alike the practice project.
Take your time to understand the concepts, so you can move on.
I'll recommend to watch the specialization of Neural Network from Andrew Ng, to deeply understand the "magic" ( linear regression, matrices, derivatives) of Neural Networks.
автор: Wei X•
I originally expected to learn more pure TF related stuff. But instead I learned Keras. Data augmentation with Keras is quite easy. Transfer learning is also easy to do if there is Keras model there already. But I do hope to learn a pure TF tutorial that are more common when you download other people's TF model and practice with your own data.
автор: Victor A N P•
Very good course and a good sequel to the first course. These courses give what we need to try our own projects. The course doesn't teach much theory, but it makes us interested and make us search and try to learn on our owns. The notebooks provided in this course, however, aren't as good as the notebooks provided in the first course.
автор: Pablo S•
Muy instructivo y activo. A uno como estudiante lo obliga a interiorizarse de verdad en los conceptos para comprender mejor las etapas que se deben implementar para el tratamiento e implementacion de una red neuronal convolucional. En general, con explicaciones claras y comprensibles puedo decir que este este un curso muy bueno.
автор: Anil K S•
This was the actual dealing with the dataset saved at local memory location rather than predefine dataset where the dealing with label and directory were ignored which learner actually face problems while learning and handling with the datasets stored at local drive. well this course actually helped for my major year project .
автор: Sagar P•
Precise and to the mark. Good brief up of the concepts. 5 stars for ease of implementation through programming assignments. Suggestion to fellow learners: Couple these courses with those by Andrew Ng, so it would be the best merger of theory + implementation. Laurence Moroney never fails individual's expectations. :)
автор: Mateus d A D P•
This course presents a more in-depth look at CNNs in comparison to the first one of this specialization. Subjects as Image Augmentation, Data Generators and others are taught about. The only thing I didn't find quite right is the final assignment. I could be wrong here, but it seems it wasn't designed properly.
автор: Ara B•
Easy to follow. a lot of examples. I was expecting at least one assignment for the final! :)
As for the convolution we never talked about DOG+SIFT or other feature extraction techniques. Also I would like to see how we can separate an object of interest from background e.g. using clustering or a video stream.
автор: ALVARO M A N•
I love this, because the instructor make the difficult easy. After ending this course, I believe I would enrolled on the other specialization, to gain a better mathematical understanding of convolutional neural networks but I'm pretty happy to learn the practical stuff, this make possible a lot of projects!
автор: Deepak V•
This course builds on the previous introductory course in the Specialisation. Not only do the four exercises provide practice towards neural network implementation, they also provide a chance to use Python for organisation and manipulation of data, pre-learning.
A fantastic and concise course over all.
автор: Aditya W•
I mainly to learn the various constructs to do various things in TensorFlow, and this course is very well constructed for it. It doesn't explain the actual mathematics though, and I don't blame it for that. It is just designed to help people learn the framework. Overall, a very satisfying experience.
автор: Jian C•
This course is a very good introduction to Tensorflow and CNN. I have taken Machine Learning theories at school and this is a very nice **programatic** supplement to my course. I think this would be even more helpful if I took it before I learn the theories. I would have been in less trouble then.
автор: Karan S•
It's amazing how far we've come in image processing. I remember using basic filters like sobel edge detector during my undergrad. And now we are here, being able to get SOTA results in just few minutes. I wonder how those Phds who were working on handcrafting filters ~2010 would have felt.
автор: Anujeet S•
This course in tensorflow specialization is a must recommended. It builds knowledge from beginners to advance very smoothly, You will be able to get a experience of how to begin coding for tensorflow also be able to understand its core layers, And learning from Laurence is always fun.
автор: Sanjay M•
Very well thought through course for Convolution Neural Networks using Tensorflow, covering some of advances topics like transfer learning, callback and review convolution layers. I already had understanding about CNN and these topics. This course shared scenarios when it is used.
автор: Thuyen T D•
The course was amazing, but the thing i don't get it is the 'sparse_categorical_crossentropy' must be use in the last exercise notebook. In the video(s) ,Laurence introduced only 'categorical_crossentropy', hope somebody could upgrade the notebook to suitable the lessons. Tks.
автор: Ozgur P•
Really good course, but recommend doing deeplerning specialization first before doing this one or doing them together. Because Andrew Ng explains really well how convolutions work, and without this background info, it will be difficult to understand the concepts in this course.
автор: Syed A A•
Really impressed by the work of the team. It is designed specially for the beginner to advance their career and be expert the emerging AI field. With the help of high quality videos and project based assignment one become expert hoe to deal and tackle real world problems.
автор: Zahid A•
What an amazing set of courses. Full practical and to point. No time wastage. Believe me if are interested in any course and DeepLearning.ai has it then blindly just enrolled in it as they have the best courses in the coursera platform. Thank you Laurence and Andrew Ng.
автор: Simon Z•
Excellent. I learned after a couple of years working with neural networks new topics and implementations. I think it would be a good idea to include also here an exercise that gets graded at the end such that we take our time and can try out if we can make things work.
автор: Abhinav S T•
The week 1 is a bit casual but where as the remaining one's are just awesome learnt a lot like how to implement a model without overfiting and learnt how to implement transfer learning and multi-class classification problem, really worthy taking up this course....!!!
автор: Waqas A•
This course is for beginners and intermediate, If you know the detail of the model layer then don't take this course. The instructor only tells the code who to add Conv, pool max layers in TensorFlow do not explain the depth of convolution and pooling layers.
автор: Shubham K•
This was a really great course for me to dive into practising the implementation of machine learning for image datasets. The instructor is really nice. I thoroughly enjoyed the course and will be taking more courses on applied machine learning from Laurence.
автор: AKSHAY K C•
The course was nicely built on the advanced topics of multi-class classification, data-augmentation, and transfer learning in Convolutional Neural Networks. Special congratulations to the instructor and his team for coming up with such a nice course.
автор: Mike B•
The course was excellent. Other than the (typical by now) Coursera code-submission issues, the course really covers a broader range of CV applications & TF capabilities than I've seen with the "get it working and move on" workflow at the day job.