As in previous lectures we peeked under the hood, but the important thing is to make our model perform really well we have to keep an eye out for lot of details.

This process requires being able to look inside the neural network as it trains and as it make predictions, find possible problems and know how to fix them

From now on will dive deep into mechanics of deep learning, things like:

• Architecture of computer vision model , NLP , tabular model.
• how to create arch that matches the needs of our particular domain.
• how to get the best possible results from the training process.
• and how to make things faster
• what changes we gotta make as the datasets change etc...

Going to make the basic application looked earlier in the Chapter 1, but will do 2 things to that :

• Make them better
• Apply them to a wider variety of types of data

Well to work out these 2 things, we gotta learn the pieces of deep learning puzzle, which includes:

• Types of layers
• Regularization methods
• optimizers
• how to put layers together into architectures
• labelling techniques etc...