Mapicx - High-Performance Neural Network Framework

Mapicx is a lightweight yet powerful neural network framework designed for both educational purposes and practical deep learning applications. Built from scratch in Python, Mapicx provides intuitive interfaces for building, training, and deploying various types of neural networks.

Key Features

• 🚀 Multi-Architecture Support
  • Artificial Neural Networks (ANN)
  • Convolutional Neural Networks (CNN)
  • Recurrent Neural Networks (RNN/LSTM)
  • Hybrid architectures
• 💻 Intuitive API
  • Keras-like syntax for easy adoption
  • Modular layer-based design
• ⚡ High Performance
  • Optimized matrix operations with NumPy
  • GPU acceleration support (coming soon)
• 📦 Extensible Design
  • Easily customizable layers and activation functions
  • Simple plugin system for custom components

Installation

pip install Mapicx

Quick Start

from Mapicx import Mapicx
from Mapicx.datasets import spiral_data
from Mapicx.optimizers import SGD

# Load dataset
X, y = spiral_data(samples=1000, classes=3)

# Create model
model = Mapicx()
model.add(2, 128, layer='Dense', activation='Relu')
model.add(128, 64, layer='Dense', activation='Relu')
model.add(64, 3, layer='Dense', activation='Softmax')

# Compile with optimizer
optimizer = SGD(_learning_rate=0.1, _decay=1e-4, momentum=0.9)
model.compile(optimizer=optimizer)

# Train the model
model.fit(X, y, epochs=1000, print_every=100)

# Make predictions
predictions = model.predict(X)

Custom Neural Architecture

model = Mapicx()
model.add(784, 256, activation='Relu')
model.add(0, 0, layer='Dropout', rate=0.4)
model.add(256, 128, activation='LSTM')  # RNN layer coming soon!
model.add(128, 10, activation='Softmax')

Key Components

Layers Dense (Fully Connected) Dropout (Regularization) Activation Functions ReLU, Leaky ReLU Sigmoid, Tanh Softmax

Optimizers

SGD (with Momentum and Decay) Adam (Upcoming) RMSprop (Upcoming) Loss Functions Categorical Crossentropy Binary Crossentropy (Upcoming) MSE (Upcoming)

Dataset	Model Architecture	Accuracy	Training Time	Loss Function	Optimizer
Spiral (3-class)	2-512-512-3	96.7%	15.2s	Categorical Crossentropy	SGD (momentum=0.9)
Linear Data	2-64-2	99.5%	8.7s	Categorical Crossentropy	SGD (decay=1e-4)
Non-Linear XOR	2-16-16-2	97.2%	22.5s	Binary Crossentropy	SGD (momentum=0.85)
Circles Dataset	2-128-128-2	92.8%	18.3s	Binary Crossentropy	SGD (decay=5e-4)

 ```
Mapicx/                      # Root directory
│
├── Mapicx/                  # Core package
│   ├── __init__.py          # Package initialization
│   ├── layer.py             # Dense, Dropout layers
│   ├── interface.py         # NeuralNetwork class
│   ├── Activation_function.py # ReLU, Softmax
│   ├── Loss_function.py     # Categorical Crossentropy
│   ├── optimizers.py        # SGD optimizer
│   ├── Data.py              # Dataset loader
│   └── mapicx.py            # High-level API
|   └── visualization.py     # Result plotting
│
│
├── .gitignore               # Ignore patterns
├── LICENSE                  # MIT License
├── README.md                # Project documentation
├── requirements.txt         # Dependencies
├── setup.py                 # Installation script
└── pyproject.toml           # Build configuration
``` 

Documentation

Explore the full documentation at mapicx.readthedocs.io

Contributing

We welcome contributions! Please see our Contribution Guidelines for details.