numpy-matrix-js

A numpy-like Matrix/Array module for Node JS and Browser JS

Installation & Set Up

NPM

In your terminal, type this command to install the module.

npm install numpy-matrix-js

After that, add this following code to the top of your file to import it.

const np = require('numpy-matrix-js')

After this, you can now use the functions available in the module.

Browser

Add this following script tag to your html file.

<script type="module">
            import np from "https://unpkg.com/[email protected]/src/html/index.js"
</script>

Documentation

Initializing

np.zeros(rows, cols)

Creates a new matrix/2d array filled with 0 If you only give 1 parameter, it creates a 1d array.

Parameters | What it is | Required -------------|----------------|---------- rows | number of rows in your array | yes cols | number of columns in your array | no

np.random.rand(rows, cols)

Creates a new matrix/2d array of random values

Parameters | What it is | Required -------------|----------------|---------- rows | number of rows in your array | yes cols | number of columns in your array | yes

Math

np.matmul(a, b)

Multiplies 2 matrices together and returns the result.

Note: Like linear algebra, the columns of the first matrix must match the rows of the second matrix.

Parameters | What it is | Required --------------|---------------|---------- a | First matrix | Yes b | Second matrix | Yes

np.add(a,b)

Adds 2 matrices together and returns the result.

Note: Like linear algebra, the dimensions of both matrices needs to match.

Parameters | What it is | Required -------------|---------------|--------- a | First matrix | Yes b | Second matrix | Yes

np.subtract(a,b)

Subtracts 2 matrices and returns the result.

Note: Like linear algebra, the dimensions of both matrices needs to match.

Parameters | What it is | Required -------------|---------------|--------- a | First matrix | Yes b | Second matrix | Yes

np.transpose(a)

Returns the transposed version of the matrix "a".

Parameters | What it is | Required -------------|---------------|--------- a | Input matrix which needs to be transposed | Yes

Built-in Math Functions

Sigmoid

np.sigmoid(x)

Returns the sigmoid value of whatever "x" value you input

Dsigmoid (Derivative of sigmoid)

np.dsigmoid(x)

Returns the dsigmoid value of whatever "x" value you input

Tanh (Hyperbolic Tangent)

np.tanh(x)

Returns the hyperbolic tangent value of whatever "x" value you input

Softmax

np.softmax(inputs)

Applies the softmax function to the inputs array and returns the new result as an array.

Relu

np.relu(x)

Applies the relu function to the value "x" and returns result.

Leaky Relu

np.leakyrelu(x)

Applies the leaky relu function to the value "x" and returns the result.

Debug

Info

let arr = [[0,0,0],[0,0,0]]
arr.info()

Automatically logs the dimensions of the array that you invoke the function on.

Table

let arr = [[0,0,0],[0,0,0]]
arr.table()

Automatically prints the array to the console in a neat format

HasNaN

let arr = [[0,0,0],[0,0,0]]
arr.hasNaN()

Checks if any of the values of the array are "NaN". If there are values with "NaN", it prints "true" to the console.

Versions

Latest Versions: 1.1.3

Stable Versions:

• 1.1.3: Fixed debug glitches. Added relu functions
• 1.1.1: Working browser support via static import