Artificial Intelligence #6 : LSTM Neural Networks with Keras (Udemy.com)

Easy to understand & well planned, cheers keep it up

This course is boring and doesn’t do a good job of conveying the information. I received this course for free, and I still want my money back!

The only reason this course is rated so highly is that the instructor tells students to not leave anything other than a 5 Star review in his course. He is clearly inflating his ratings.

Buyer be warned...

Great course I really like way of teaching. Completely learn the basic concept you are teaching . Nice job.

The instructor has very poor language and very poor knowledge. He does not explain the material behind the scene.

Axel,

First of all thank you. I'm not new in machine learning but yes in Tensorflow and Keras. I would like to suggest you can add an introduction explaining briefly Keras' API, for a better understanding. The code is clean and works as expected, this is a hands on training where it is supposed the student has the experience required. One additional chapter could be "how to improve" all the previous examples and the last in particular (wind) where changing epochs and LSTM's parameters is not enough.

José Valerio

Ripasso di keras e concetti ripetuti su come presentare i dati alla rete,

qualche imprecisione e nessun riferimento al concetto di stato e contesto ne alle loro impostazioni nel training (stateful). Detto ciò è un ripasso del codice :)

dragging

Dear Dr. Sobhan,

Thank you for the great course, really I learned a lot of things from it. I hope you will give more courses of LSTM and Neural Network. Thank you.

Covers how to build a sequential LSTM network and how to train it. Shows how to preprocess the data using scipy. Seems to repeat this for 3 data sets.

Quite a slow drawn out process, however perhaps that is good for total beginners. Perhaps a shorter course covering the same material would be better.

this is not a prediction, just anti-aliasing! you spent my 15 minutes.

some code lines are not really explained why they must exist for the code to work. the datastructures are of as much importance as understanding what the neural network does internally. the input datastructures to the network must be explained in more detail.

For me the tempo is good and the quality of the example taken also good.

I have seen my questions asked by others, and answered by you, in the comments. And they were really answered very extensively and well, therefore I give you 1 star more.

Your explanation was vague at many places and you didn't explained thoroughly each codes. I liked the examples on LSTM but I am still hardly able to understand the code.