Unit 2: Explaining the "residual learning"

[0xD4rky]

I would like to explain the residual learning, introduced in the official paper, in depth.

I want to explain how learning (h(x)-x) is easier for the model rather than learning h(x) (where h(x) is the function that maps the input and output of the stacked layer).

Hence, allow me to raise a PR for updating the docs and you review the changes!