Metadata-Version: 2.1
Name: pynetworklib
Version: 1.0.1
Summary: A simple library for makeing dense multilayer neural networks
Author: Joseph Bronsten
Author-email: josephbronsten.dev_contact@gmail.com
Description-Content-Type: text/markdown
Requires-Dist: numpy

# PyNetworkLib

### This is an INDEV version of a multi-layer perception network written in Python

### It is designed to be used as a library, it has a network module that provides trainable dense multilayer perceptron networks, a configuration module that provides a way to store parameters for training and optimization, and an optimization module that creates an optimized network structure and hyperparameters for a given dataset

## Installation of the library

```bash
pip install pynetworklib
```

## Example of how to use this library

```python
from network import network
from network import dense
from network import activation_functions
from network import mse
import numpy as np

# Create a network with a struct of 3 layers of size 2, 3, and 1
input_size = 2
output_size = 1
network_xor = network.Network([dense.Dense(2,3), activation_functions.Tanh(), 
                           dense.Dense(3,1), activation_functions.Tanh()], 
                           input_size, output_size)

# Create a data set, this example uses the xor operation as the training output
x_train = [
    np.array([1,1]),
    np.array([0,1]),
    np.array([1,0]),
    np.array([0,0])
    ]
y_train = [0,1,1,0]

network_xor.fit(mse.MeanSquaredError(), x_train, y_train, epoachs = 10, verbose = True)


```
