Deep Learning: Recurrent Neural Networks in Python (Udemy.com)

I loved the approach on how core ideas were presented before starting with RNN (time-series , forecasting and a linear version of RNN) and how much an ANN/FFNN could do without having to use RNNs. This gave a good picture of why a RNN was needed at the first place. And, not to mention the response on Q&A from the instructor is timely and very helpful. In some of the notebook there were little code issues, but, because the concepts were outlined so, well that were not hard to self resolve so, I don't mind that. Because the concepts are outlined well, I don't feel very un-comfortable when I see code now - whether it is Tensorflow or Pytorch. Lastly, at this time as I have taken multiple courses from LazyProgrammer, I am very thankful as it is helping me very much to understand deep learning and AI in general.

The lazy programmer is a very good teacher, probably way better than him being a Machine Learning Engineer/Pratitioner.He must have invested a considerable amount of time to teach effectively to students coming from a wide variety of backgrounds. Very well done! Thanks.

This course was out of this world, I've ever seen before! Especially, section11-13. Off course, I can't say, I can get it everything you were taught me, but at least, I can see how to build my machine-learning ability or learning path. Your golden rule "All data is the same." will be my golden rule as well. I will always keep in mind this phrase!

This course has provided me with the best comprehension of RNNs and LSTMs I've ever had. The diagrams used in other more high-level intuition-based courses did more harm than good by causing more confusion than elucidation. Once you appreciate the basic math behind how RNNs work, you realize they're not all that scary.

Overall, the course was absolutely stunning. Appreciate the teaching style which included introducing each topic from a high level overview, building from simplicity and encouragement of writing code. Found the course progresses rapidly to an advanced level, but you can take your time and go your own pace. It's not too difficult to follow along.

Was exactly what I needed to get along with the Tensorflow 2 framework. I took the instructor's prerequisite course to get the basics of Numpy, Pandas, Matplotlib, ... and moved on to more advanced ML topics. After 15 years of using SAS it's really refreshing to use Python. :) Thank you.

Short version: Decent course. Learned about RNN's and LSTM's. Some theory, and some hands on experience. I would recommend, and I would take another Lazy Programming course. TLDNR: This is my first Lazy Programmer course, but not my first machine learning course. The instructor seems very knowledgeable and experienced. He has a good sense of humor, and does not take himself too seriously. I find this to be the mark of an individual confident in themselves. The course did get me familiar with RNN's and LSTM's. You have to have initiative for this course, and it sounds like this is the case for most of Lazy Programmer's courses. There were not any specific exercises or capstone projects given. It is up to you to try to replicate code demonstrations yourself to test your understanding. Overall, the course was worth my time and I would take another Lazy Programmer course. I would like to suggest a few things for improvement. First, not so much fluff. I feel like only about half of the stated course duration's time was actually spent on RNN's and LSTM's. Second, I fee like the course could be organized a little better. For example, effective learning strategies and coding for beginners seems better situated at the beginning of the course. Students that actually need that info might not find it until after the course is completed. And there is something unnatural about going to the end of a course to see how the instructor thinks you should approach the course. Most students that do not need that info will just skip it, if it is placed at the beginning of the course. Also, in the course description, it would be good to link to your website to explain what order to take your courses in. Having to hunt around for the video to describe course order that you said was in all your courses was not helpful. Third, this course was pretty good at being stand alone, but at times you were depending on people to have watched a previous course lector. It was not so much a matter of needing a technical skill from another course, but rather requiring specific parts of a conversation to be recalled. This usually occurred when the video appeared to be cut from another course. It did not flow so well. Fourth, more explanation would be helpful. I really enjoyed the theory, algorithm decomposition, and code implementation. I think there are even more opportunities to do this. And finally, I feel it is incumbent on instructors to create ample learning opportunities for students. Well thought out, explicit exercises that are just challenging enough are where people learn the most. Like you say, if yo cannot implement it, you do not understand it. I would encourage you to create go/no-go exercise that force the student to test there understand at critical junctions. Otherwise, the video becomes less interactive and more book-like. Thanks for the course. You are probably the fifth or sixth lecturer/author I have learned from, and I think you have a lot to offer. Stop being so lazy! Just kidding.

Overall the course content is good. It covers the basics of RNNs with coding examples explained well. I would recommend this course to anyone who is interested in learning about RNNs. A couple of things that could be added to the course: 1. A video lecture dedicated to how backpropagation works in training RNNs. 2. A video lecture explaining mathematically how LSTMs solves (mitigates) the vanishing gradient problems.

The course was pretty good so far, the instructor stresses on calculus behind the practical tools like tensorflow/keras. I believe this is more helpful for the engineers to have solid idea behind the simple functions. Just to add little thing for better understanding, in examples it's better to be explained that why some loss function is chosen with the data is used.

I've been asked a review after the initial few lectures, so decided to rate the knowledge shared and gained so far. The lecture's are explained very well and can grasp the underlying mechanics of deep learning much easier. Finished the course and here goes my review, Neural networks were much better explained for me than anywhere else. And the contents are top end. Learnt a lot and it gives us confidence to go ahead and experiment and build further neural networks ourselves.