Releases · thieu1995/WaveletML · GitHub

04 Jun 18:11

thieu1995

v0.2.0 Latest

Latest

[0.2.0] – Major Update

🔧 Enhancements

Improved setup.py for more robust and maintainable package management.
Upgraded mealpy dependency to v3.0.2 for better performance and compatibility.
Updated GitHub Actions workflows for testing and publishing.

🧠 Core Module Updates

Enhanced BaseMhaWnnModel:
- Added support for additional parameters in the __init__() method.
Refactored:
- data_preparer and data_scaler modules with improved docstrings and minor internal adjustments.

📚 Documentation & Testing

Updated example scripts for clarity and consistency.
Improved unit tests across modules.
Revised and expanded documentation.

Assets 2

25 May 06:42

thieu1995

v0.1.0

[0.1.0] - Initial Release

📁 Project Structure

Added essential project files: CODE_OF_CONDUCT.md, MANIFEST.in, LICENSE, CITATION.cff, and requirements.txt
Added structured layout for examples/, tests/, and docs/ (documentation site)

🧰 Helpers Module

Added helpers package:
- verifier: input and parameter validation
- evaluator: evaluation metrics
- data_scaler: feature normalization
- data_preparer: dataset scaling and splitting
- callbacks: custom callback functionality
- wavelet_funcs: wavelet function definitions and management
- wavelet_layers: PyTorch-based wavelet layer implementations

🧠 Models Package

Added models package:
- base_model.py: defines BaseModel for consistent design and logic reuse
- custom_wnn: custom wavelet neural network implementations (4 types)
- gd_wnn: fully gradient-descent-based WNNs:
  - GdWnnClassifier: for classification tasks
  - GdWnnRegressor: for regression tasks
- mha_wnn: fully metaheuristic-optimized WNNs:
  - MhaWnnClassifier: for classification
  - MhaWnnRegressor: for regression

Assets 2