atlas-aerospace-yt / vector

A python vector library

Vector

A python vector library

Contents

• Why vector
• Usage
• Installation

Why vector

Vector is a mathematical library for python. In my search to understand artificial intelligence (AI) I strived to make my own neural network entirely from scratch. To do this I began with numpy – a python library which is the go-to for mathematical things within python. As I started to make this program, I realised that I wanted more control over the objects and wanted to use them in such a way that would make sense for AI. So, I made my own library – vector – a simple vector library in python which holds the basic activation functions for neural networks, random generation, and of course a Vector object which is to be used in a way similar to np.array.

Usage

To import the library:

from vector import Vector
from vector import activation
from vector import random

To define a vector object, a = Vector([1, 0, 0, 1]) or a = Vector(1). This class definition takes in either a list, tuple, int, or float.
To generate a random vector, a = random.random_vector(length, lower_bound, upper_bound), lower_bound and upper_bound are optional arguments and by default are -0.1 and 0.1 respectively.
The vector library has 2 activation functions and the 2 derivative functions of those functions:

activation.sigmoig(vec)
activation.sigmoid_prime(vec)
activation.linear(vec)
activation.linear_prime(vec)

In the example above, vec is either an integer, float, or Vector. sigmoid is: 1 / (1 + e ^ -x) and linear is: 0.5 * x. The _prime in the function name means derivative.

Installation

To install this library using pip, you need to type pip install vector-ai-ml.