New Fast.ai Course: A Code-First Introduction to Natural Language Processing

Maybe read their code and it's associated comments would help you? One of the benefits of using notebooks over a traditional IDE is that code is segmented into blocks. For me it makes it easier to read. it's like "ok, this cell will do this to our data, then this next cell will do this to our data". It's been more than a year since I took their class, and what I did to help with comprehension of the videos was to take notes in the individual cells while following the videos.

For at least some of the lessons there are pretty comprehensive notes / transcripts linked in the column on the right.

E.g. for https://course.fast.ai/videos/?lesson=2 there is https://github.com/hiromis/notes/blob/master/Lesson2.md

What is your skillset? What language?

(My own writings)

Deep Learning from Scratch to GPU at https://dragan.rocks

Deep Learning for Programmers (book in progress) at https://aiprobook.com/deep-learning-for-programmers

Project idea: use the NLP skills taught in this course to build an automated system that converts the course's videos into a written tutorial.

I have the same problem with video courses. The moment I start to sit and watch, I feel drowsy. It's strange, because I never had it so bad in real life lectures. Did anyone experience the same and find a way to fix themselves?

My gf worked for Coursera for a while so I tried some of their courses. The big win for me was that every one I tried included a full transcript. So my use model was to simply read the transcripts and if I didn't understand something I clicked on the text which would jump to that point in the video.

But in general for the stuff I want to learn I prefer to read a book, which is harder to find for really recent stuff like this.

Yes.

An Introduction to Statistical Learning https://github.com/tpn/pdfs/blob/master/An%20Introduction%20...

Deep Learning https://books.google.com/books/about/Deep_Learning.html?id=o...

Machine Learning Mastery books https://machinelearningmastery.com/products/

Convolutional Neural Networks from the Ground Up https://towardsdatascience.com/convolutional-neural-networks...

Transformers https://medium.com/inside-machine-learning/what-is-a-transfo...

For deep learning / neural networks, I think these are the best balance of theory and practice (not to mention great visualizations): http://cs231n.github.io

More on the intuition/theory side for NNs, but Michael Nielsen’s book is highly recommended: http://neuralnetworksanddeeplearning.com

For NLP, you can try this extremely slick interactive course on spaCy (highly recommended Swiss Army knife library for NLP) from package author Ines Montani: https://course.spacy.io

The Keras tutorials are solid, here’s one on seq2seq models: https://blog.keras.io/a-ten-minute-introduction-to-sequence-...

For more cutting edge NLP, it looks like the fast.ai course covers the Transformer model (the basis of BERT, GPT-2, et al. This is a great overview of that architecture: http://nlp.seas.harvard.edu/2018/04/03/attention.html

And PyTorch implementations of Transformer models can be found here: https://github.com/huggingface/pytorch-pretrained-BERT/blob/...

LMK if anyone wants more, I have a learning resource hoarding problem.

https://www.deeplearningbook.org/ is also great