Skip to content

CHANGELOG.md

Latest commit

 

History

History
140 lines (100 loc) · 4.18 KB

CHANGELOG.md

File metadata and controls

140 lines (100 loc) · 4.18 KB

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.

Unreleased

Planned

  • Video sequence support for temporal tasks
  • Integration with Weights & Biases for experiment tracking
  • TensorRT optimization for faster inference
  • Docker images for easy deployment
  • Additional model architectures (DETR, YOLOv8, etc.)
  • Mobile deployment guide (CoreML, TFLite)

[2.10] - 2025-01-18

Add GitHub Workflows

  • Added policies for PRs and CI for testing
  • Added bib files for referencing

2.0.0 - 2025-01-15

Major Rewrite

Complete rewrite of the original 2019 toolkit with modern PyTorch and best practices.

Added

  • Core Components

    • Native PyTorch Dataset class (VisDroneDataset) with automatic filtering
    • Support for 4 detection models: Faster R-CNN (ResNet50, MobileNet), FCOS, RetinaNet
    • Model factory with get_model() function
    • Comprehensive visualization utilities
    • Format converters for COCO and YOLO (not just VOC)

  • CLI Tools

    • train.py - Complete training pipeline with AMP support
    • inference.py - Batch inference on images/videos
    • webcam_demo.py - Real-time webcam detection demo
    • evaluate.py - Model evaluation with metrics
    • convert_annotations.py - Format conversion utility

  • Training Features

    • Automatic mixed precision (AMP) for 2x faster training
    • Learning rate scheduling (StepLR)
    • Checkpointing and resuming
    • Training curve visualization
    • Validation during training

  • Documentation

    • Comprehensive README with examples
    • Installation guide (INSTALL.md)
    • Quick reference guide (QUICKSTART.md)
    • Contributing guidelines (CONTRIBUTING.md)
    • Code of Conduct (CODE_OF_CONDUCT.md)
    • Security policy (SECURITY.md)
    • Test documentation

  • Development

    • 66 unit tests with pytest
    • Test coverage >80%
    • Pre-commit hooks for code quality
    • Makefile for common tasks
    • Type hints throughout codebase
    • Modern packaging with pyproject.toml

  • Configuration

    • YAML configuration files for all models
    • Virtualenv-first installation
    • Requirements files for different use cases
    • .gitignore for Python projects

Changed

  • Framework: Migrated from TensorFlow 1.x to PyTorch 2.0+
  • Models: Replaced custom implementations with torchvision models
  • Format: Moved from PASCAL VOC only to COCO and YOLO support
  • Installation: Simplified with virtualenv instead of conda
  • Structure: Proper Python package with CLI entry points

Improved

  • Performance: 2-3x faster training with AMP
  • Usability: Simple CLI tools instead of complex scripts
  • Maintainability: Comprehensive tests and documentation
  • Code Quality: Type hints, linting, formatting with Black
  • Flexibility: Easy to extend with new models and features

Removed

  • TensorFlow dependencies
  • PASCAL VOC as primary format (still available if needed)
  • Legacy training scripts
  • Conda environment (replaced with virtualenv)

1.0.0 - 2019-XX-XX

Initial Release (Legacy)

  • Basic VisDrone annotation conversion to PASCAL VOC format
  • TensorFlow 1.x implementation
  • Faster R-CNN with Inception V3
  • Simple conversion script (convertVis_to_xml.py)
  • Basic visualization
  • Training script for static images

Version Numbering

We use Semantic Versioning:

  • MAJOR version for incompatible API changes
  • MINOR version for new functionality in a backwards compatible manner
  • PATCH version for backwards compatible bug fixes

Types of Changes

  • Added - New features
  • Changed - Changes in existing functionality
  • Deprecated - Soon-to-be removed features
  • Removed - Removed features
  • Fixed - Bug fixes
  • Security - Security vulnerability fixes

Links