Skip to content

v0.4.0

Latest
Latest
Compare
Choose a tag to compare
@acsenrafilho acsenrafilho released this 30 Nov 21:56
· 4 commits to main since this release
v0.4.0
a5cb8c9

Release Notes: Version v0.4.0

Overview

This release introduces comprehensive updates, including new machine learning modules, expanded documentation, and critical bug fixes. The addition of image classification support, Kaggle integration, and adjustments to key machine learning models significantly enhance the repository's functionality.

🚀 New Features

Machine Learning Enhancements

  • Added SmallXception Architecture:
    A lightweight model for binary image classification tasks.
    Commit: b2818e4, b661a0e

  • Introduced ImageClassificationTrainer Class:
    Facilitates training and evaluation workflows for image classification.
    Commit: 180ea0e

  • Added ml_models and ml_trainers Modules:
    Comprehensive modules for defining and training machine learning models.
    Commit: 8c1b804, 4577d5f

  • Kaggle Integration:

    • New kagglehub dependency for downloading datasets and pre-trained models.
      Commit: 77c832f
    • Included default addresses for Cucaracha pre-trained models.
      Commit: 4c9c711
    • Module for handling Kaggle API calls.
      Commit: 1ce6e56

  • Expanded Testing:

    • Tests for verify_image_compatibility and ml_models modules.
      Commits: 03754ac, 0d92f22

Dependency and Compatibility Updates

  • TensorFlow 2.16.2 defined as the required version.
    Commit: 63cd4c9
  • Removed Windows compatibility; supports Linux/Mac only.
    Commit: 963fa74

🐛 Bug Fixes

  • Dataset Loading Fix:
    Adjusted load_cucaracha_dataset to rearrange folder structures for compatibility with ml_trainers.
    Commit: 515d01c

  • Model Improvements:
    Fixed lint issues with the SmallXception model and __str__ method for the ModelArchitec interface.
    Commits: 799474c, 436f352

  • Various Fixes:

    • Resolved int to float assertion issues.
      Commit: 2fee044
    • Fixes for TensorFlow training runner stability.
      Commit: 279d75b

📝 Documentation Updates

  • Getting Started Section Added:
    Comprehensive guide for new users.
    Commit: eb809a8

  • New Documentation Sections:
    Added details about ml_models and ml_trainers modules.
    Commit: f727042

  • Documentation Configuration:
    Added mkdocs.yml for structured documentation generation.
    Commit: a5cb8c9

🛠️ Enhancements

  • Organized the MLPattern interface for general ML implementations.
    Commit: c355638

  • Improved code coverage and file organization within the repository.
    Commit: 335a85c, 104ae7f

  • Updated .gitignore to exclude .keras files and organized_dataset.
    Commits: 45bdcfb, f381dae

  • Improved dataset samples for more robust testing.
    Commit: c9c8636

🔒 Issues Resolved

  • #17: Add ML training support.
  • #12: Define ML model and trainer modules.
  • #8: Add Kaggle integration for datasets and models.
  • #7: Support for image classification tasks.
  • #6: TensorFlow training runner fixes.

Commit Summary

Commit Type Description
a5cb8c9 DOC Add mkdocs.yml and getting_started.md section
799474c BUG Fix lint on SmallXception model
279d75b BUG Fix TensorFlow training runner
b2818e4 ENH Add SmallXception ML model
1ce6e56 ENH Kaggle API integration
eb809a8 DOC Added Getting Started guide
515d01c BUG Fix dataset loading issues
b661a0e ENH Adjust SmallXception to binary classification
f727042 DOC Add sections for ML modules
8c1b804 ENH Define ml_models and ml_trainers modules
77c832f ENH Add kagglehub dependency
2fee044 BUG Fix int-to-float assertion
335a85c ENH Improved code coverage
63cd4c9 ENH Defined TensorFlow 2.16.2 compatibility

Upgrade Instructions

  1. Update your package to the latest version:

    pip install cucaracha --upgrade

  2. Follow the Getting Started Guide for an overview of the new features.

We encourage users to explore the new ML capabilities and provide feedback through GitHub issues. Thank you for supporting

Assets 2
Loading