Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.1.0] - 2026-06-25

Initial release of the gate_learner neural network engine in Python.
Custom Multilayer Perceptron (MLP) implementation with support for arbitrary hidden layer architectures.
Sigmoid and ReLU activation functions with Xavier/Glorot and He Uniform weight initialisation.
Adam Optimizer with L2 regularisation, momentum, and bias correction.
Pre-allocated internal buffers for activations, deltas, and gradients to avoid repetitive memory allocations during training.
Dynamic dataset generator for $N$-input OR and XOR logic gates.
Early stopping mechanisms based on convergence (Loss < 0.01) and stalling (Loss change < 1e-6 over 50 epochs).
JSON serialisation and deserialisation for models and training histories.
Dual-axis plotting of training loss and accuracy curves (gated under the visualise extra feature using matplotlib).
Comprehensive unit tests, integration tests, and performance benchmarks using pytest and pytest-benchmark.