Deep Learning Resources for 2019

Machine and Deep Learning move so fast that from year to year resources become easily outdated and new tools and libraries emerge. Here is a guide for of tutorial and resources for 2019


First of all I recommend the excellent course by Jeremy Howard at San Francisco University which is free to follow online on their website. This is the new course in 2019:

There is a courses in progress and it will be uploaded to the website as it finishes.

I watched all their 2018 videos and they are fantastic and still worth to watch:

Fastai has a very lively forum For instance lets see a suggested resource list for learning Python here.


Visit and get a feeling for competition and what will be expected from ML and DL practitioners.

You can learn Pandas on Kaggle with interactive jupyter notebooks. They have quite a number of very interesting tutorials.

Andrew Ng on Coursera

Both courses can be watched free if you enroll as auditor. Please see my article about “How to audit courses on Coursera” and how to benefit of the MOOC offering. The oldest of his courses,

has been my introduction to Machine Learning. When I first watched this I felt like superman. I used to think ML was beyond my reach (and it still is) but I could develop an intuition for it and I got greatly inspired. Terrific course but now outdated. I believe it has been made around 2012 and they still use octave software for assignments. I would still watch the courses but would recommend to start with:

In this specialisation in 5 in 5 coursecourses he uses Python and Jupyter notebooks for the assignments and it goes really indeep about forward and back-propagation. Great for your intuition. Not always easy. Some assignments are available for auditors. Probably pay for the course is good value for money, even if you are not interested in the certificates, the assignments are great value to learn. As a curiosity the seminal course by Geoffrey Hinton, “Neural Networks for Machine Learning” on Coursera is not offered anymore but still to watch on youtube.

Stanford CS231n

The CS231n: Convolutional Neural Networks for Visual Recognition course at Stanford is quite popular. The 2017 videos are on youtube.

It is good to watch but assignments and resource websites are already a bit outdated. Great material though, waiting to watch the 2018 videos when they become available.

A course in Algebra

taught by Rachel Thomas (one of the co-founder of fastai) at the university of San Francisco is a series of videos about that math knowledge you migt need for deep learning. The Python notebooks are on her GitHub

A very practical hands-on tutorials

He has a very practical blog with lots of tips. The good part is that many of the articles get you up and running in a minimal amount of time. He has written three books on the subject and they are quite expensive. The website and blogs are really promotion for his books but I like his approach and he really always replies to comments ad questions.


It is better not to spend too much time on the theory, the practical side of AI should be prioritise, Jeremy Howard talks of a top-down approach and many agree with him. Spend time coding and getting the knowledge you need when you need it. However as a reference I found these books to be a good companion to your journey.

Read some good blogs

And if this was not enough for you, head to another incredibly detailed list published on Kaggle :Awesome Deep Learning Basics and Resources by Arunkumar V Ramanan

That’s it for now. If I missed something let me know in the comments!

If you liked this post, you can share it with your followers or follow me on Twitter!