Modern Natural Language Processing in Python

It was a good match, although some code snippets are old, most work as intended and are clear enough that you can work around it.

This is the only tutorial in Udemy in which we can find the English to French translation model. Very nice explanation about the Transformer model and all its detailed implementation.

The course included detailed explanations of classification and translation, and I was able to accomplish the same steps using Jupyter instead of Google Colab.

Really good coures. It doesn't take longer that it needs so you don't waste energy following redundant explanations. Right to the point!

it was not been told that we need a background in deep learning and TensorFlow. so I assume that the lecturer will explain his steps, but he basically just write it.

This is an advanced course and covers recent concepts like Transformer, Positional Encoding, Encoder, Decoder etc. It requires some prior knowledge of CNN, RNN, LSTM, seq-to-seq learning. The course also covers building a Transformer from scratch using the "Attention is all you need" research paper as basis with English to French translation as a use case. However, the course contents are very small and explanations are "verbal". It should be extended with more visuals and use of white board to convey the concepts well. Also, content should be simplified by using some examples while explaining the concepts. I found Hedu Math of Intelligence YT videos and Illustrated Transformer blog by Jay Alammar helpful to understand these concepts. The coding part should also be simplified by explaining the logic & design, and showing running samples of expected output like matrix/vector shapes, sample data etc. It becomes very hard to keep up with the intuitions and code implementations. Overall, the course covers an advanced topic which is hard to apprehend initially and requires some prior knowledge; but it is also a great opportunity to simplify the concepts given that there are not many resources available for the same.

The lessons are very understandable. and the content is broken down that makes it fully digestible. Enjoying the content.

A good course on cutting edge topics. Some parts of the transformer were quite heavy and difficult to follow. I know it is difficult to visualize the internals of such a complex system, but I would have preferred visualization of the dimensions of each output of the different steps. It would have made it easier to follow.

It is a good course with detailed explanation of neural networks, but the content seems smaller. It is just 2 projects and their explanation

Intuion part for transformer is a little hard to follow. It should be better to use real sentence to walk throught the modle. Coding is well explained.

It's not too verbose and straight to the point. It gives you experience with state of the art NLP models as of May 2021, when I wrote this review.

This course is a an excellent refresher of the CNN module in Deep Learning A-Z. The application of CNNs to NLP is something new for me, so very valuable to learn.

Yes, However I think the explanation for the coding stuff could have been better. The theory part was good though.

Lots of talking, but few examples / explanations. Author is definitely not a teacher. Author needs a typing course.

It is a great material to understand the Transformer Architecture. Then you can play with the code, listen it again, read the paper and learn a lot.