code_first_ml

Learn ML by building a framework!

This is a series of notebooks explaining the processes necessary to train neural networks, with a focus on practical implemention in Python code. The work is based on the minai framework built in fast.ai's excellent 'From Deep Learning Foundations to Stable Diffusion' course.

The notebooks cover techniques including:

• Matrix multiplication

• Stochastic gradient descent

• Attention

• Activations