The spelled

micrograd overview

starting the core Value object of micrograd and its visualization

derivative of a function with multiple inputs

derivative of a simple function with one input

implementing the backward function for each operation

fixing a backprop bug when one node is used multiple times

implementing the backward function for a whole expression graph

preview of a single optimization step

counting bigrams in a 2D torch tensor ("training the model")

breaking up a tanh, exercising with more operations