Build a simple neural network from scratch, only using #numpy

In this video, I have tried to explain how to implement a basic linear softmax classifier from scratch, i.e. only using numpy.

This video is mathematically heavy, so if you are unfamiliar with linear algebra and calculus concepts, I would highly recommend watching my linear algebra playlist. The calculus here is pretty easy.

after learning linear algebra neural networks would be much easier to understand, speaking from experience.

chapters:
0:00 intro
0:30 Agenda
2:08 Theory
2:59 MLP | multilayer perceptron
7:20 Concept of One hot vector
9:00 Loss function
10:49 The gradient matrix
13:33 Decomposition of the gradient matrix
16:45 Computing the diagonal matrix
20:00 Jumping into the Code
21:28 Your Own neural net
27:06 Let's visualize the training process

#thepylama #machinelearning #python #math #deeplearning #implementation #softmax #neuralnetwork #backpropagation #numpy