minicnn 0.2.0

Minimal, well-documented CNN models for image classification. Great for beginners!

MiniCNN

MiniCNN is a minimal, well-documented Python package with beginner-friendly Convolutional Neural Network (CNN) models using PyTorch.
Great for learning, tutorials, and fast prototyping.

Installation

pip install MiniCNN

Usage

import torch

from MiniCNN import SimpleCNN, TinyCNN

Example: Create a SimpleCNN model for MNIST (10 classes)

model = SimpleCNN(num_classes=10)
print(model)

Random input (batch=1, channel=1, size=28x28)

x = torch.randn(1, 1, 28, 28)
y = model(x)
print(y.shape) # torch.Size()

Features

• SimpleCNN: Standard small CNN, 2 conv layers + 2 FC layers.
• TinyCNN: Extremely small model for quick tests.

Requirements

• Python 3.8+
• PyTorch >= 2.0.0

License

MIT License