Deep Scratch - A deep learning lib from scratch

(or from numpy by the way)

Main purposes:

It became easier to understand (for me, of course) the following models:

• MLP,
• ConvNet,
• DeConvNet,
• RBM,

When I started to implement

these models myself, everything became brighter, more intuitive, sometimes obvius - compared to the mathematical formulas representing a clear idea

Currently my primary goal

Is to make some sense of what are those neurons are 'thinking', inside a large net, hoping, that someday I will be able to retrain useful information from living organisms' inner state based on my current study

Generally speaking, understanding could be called some kind of visualisation...