Metadata-Version: 2.4
Name: PyAiNetwork
Version: 0.2.2
Summary: A simple Python library for creating neural networks
Author: eyes-studio
License: MIT
Requires-Python: >=3.8
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy
Dynamic: license-file

# PyAiNetwork

A lightweight, open-source Python library for building and training neural networks with only a few lines of code.

> Create AI models without writing hundreds of lines for neurons, layers, and weights.

---

## Installation

```bash
pip install PyAiNetwork
```

---

## Features

- 🚀 Beginner-friendly API
- 🧠 Standard `Network`
- ⚡ Advanced `ProfNetwork`
- 🔥 GELU activation
- 🔥 ReLU activation
- 🔥 Sigmoid activation
- 📈 Built-in training
- 🎯 Custom network architectures
- 🪶 Lightweight & fast
- 🐍 Pure Python
- 💚 Open Source

---

# Quick Start

```python
from PyAiNetwork import Network

net = Network(
    2,      # Input neurons
    2,      # Hidden layers
    4,      # Neurons per hidden layer
    1       # Output neurons
)

result = net.forward([0.5, 1.0])

print(result)
```

---

# Training

```python
from PyAiNetwork import Network

net = Network(2, 1, 4, 1)

for _ in range(100):
    net.train(
        [1, 0],
        [1]
    )

print(net.forward([1, 0]))
```

---

# Activation Functions

PyAiNetwork supports multiple activation functions.

### GELU

```python
Network(..., activition="gelu")
```

### ReLU

```python
Network(..., activition="relu")
```

### Sigmoid

```python
Network(..., activition="sigmoid")
```

---

# Network

`Network` creates a neural network where every hidden layer contains the same number of neurons.

```python
Network(
    input_neurons,
    hidden_layers,
    neurons_per_layer,
    output_neurons,
    activition="gelu"
)
```

Example

```python
net = Network(
    4,
    3,
    16,
    2
)
```

Architecture

```
4 → 16 → 16 → 16 → 2
```

---

# ProfNetwork

`ProfNetwork` allows every hidden layer to have a different number of neurons.

```python
from PyAiNetwork import ProfNetwork

net = ProfNetwork(
    2,
    [8, 16, 8],
    1
)
```

Architecture

```
2 → 8 → 16 → 8 → 1
```

Another example

```python
net = ProfNetwork(
    3,
    [32, 64, 64, 32],
    5
)
```

---

# Which one should I use?

| Class | Best for |
|--------|----------|
| Network | Simple projects and learning |
| ProfNetwork | Custom architectures and larger AI models |

---

# Roadmap

Planned features for future releases

- Adam Optimizer
- Save / Load models
- Batch training
- More activation functions
- More loss functions
- Better performance
- More neural network types
- GPU support (planned)

---

# About Eyes Studio

PyAiNetwork is developed by **Eyes Studio**, an independent software developer focused on AI technologies, developer tools, and open-source software.

---

# License

MIT License

Copyright © 2026 Eyes Studio

---

# GitHub

https://github.com/eyes-studio/PyAiNetwork
