This is the best way to get into deep learning: Do both FastAi, and also do Deeplearning.ai. They take opposite approaches that compliment each other perfectly

This is the best way to get into deep learning: Do both FastAi, and also do Deeplearning.ai. They take opposite approaches that compliment each other perfectly. Fastai starts with the practical and doubles back to help you understand the technical. So in lesson 1 of fastai you're doing practical work, and half way through the course you start learning about taking derivatives of loss functions. Deeplearning.ai on the other hand in lesson 1 you're learning about taking derivatives of loss functions, and you do practical things with it starting about half way through the course. They both are absolutely fabulous for their own reasons, but the two together really cover each other's strengths and weaknesses in an amazing way.