Skip to content

Latest commit

 

History

History
222 lines (138 loc) · 7.45 KB

CHANGELOG.md

File metadata and controls

222 lines (138 loc) · 7.45 KB

SPDX-FileCopyrightText: 2025 Doleus contributors

SPDX-License-Identifier: Apache-2.0

CHANGELOG

v0.2.4 (2025-12-16)

Features

  • Added AUPRC (Area Under Precision-Recall Curve) metric support (a7ff4b8)

Testing

  • Added AUPRC tests for binary, multiclass, and multilabel classification (3cdf6b0)

Bug Fixes

  • Replaced numpy with PyTorch in TPR/FPR functions for better performance (c8013f4)

Build

  • Removed torchvision from core dependencies (1b09c27)

Documentation

  • Added CPU-only PyTorch installation instructions (0df9619)

v0.2.2 (2025-10-04)

Build

  • Loosened dependency constraints to support wider range of PyTorch versions (torch>=2.0.0, torchvision>=0.15.0)
    • Fixes compatibility issues with Google Colab and other environments using newer PyTorch versions
    • All tests pass with loosened constraints

v0.2.1 (2025-10-04) - YANKED

Note

  • This version was yanked from PyPI due to untested dependency constraints
  • Users should use v0.2.2 or later

v0.2.0 (2025-10-04)

Build

  • Replaced opencv-python with opencv-python-headless (5f6b51c)

Features

  • Support complex slicing conditions with AND/OR logic (4e881f5)

  • Added slicing operators: in, not_in, between (75fa504)

  • Added not_between operator (a4d3442)

Documentation

  • Added docs for new slicing methods (50f77d0)

Testing

  • Added tests for complex slicing conditions (AND/OR operators) (4e881f5)

  • Added tests for automatically created metadata (65c10c3)

  • Test for not_between operator (7cec696)

v0.1.1 (2025-03-10)

Bug Fixes

Chores

Documentation

Refactoring

  • Changed docstrings (45bf991)

  • Cleaned up codebase (e9173ce)

  • Complete refactoring (ce589c5)

  • Refactored dataset class (4d36e9a)

  • Refactored metric and dataset class (c2a259d)

  • Refactored slicing and metadata addition functions (33a76d8)

  • Remove redundant files (690cb7c)

  • Removed implicit caching and upload functions (40f4986)

  • Removed redundant files (c8dccc0)

  • Removed unnecessairy files (c27ab5f)

  • Restructured metric functions (8c9dadb)

  • Simplified annotation class (4cbd92a)

Testing

  • Adapted tests for dataset class and created conftest file (e040216)

  • Additional test for dataset (a1f6a83)

v0.1.0 (2025-01-09)

Bug Fixes

  • Corrected handling of predictions (53d4d03)

Continuous Integration

Features

  • Add run method to checks (6645fcb)

  • Added tiny imagenet demo (7540c7c)

Refactoring

  • Added code to convert 0-d scalar tensors to 1-d tensors to adhere to annotations.py. dataset.py is now only accepting tensors, no lists. (663b88a)

  • Added more informative error messages (5ee2b45)

  • Added more validation checks and grouped them in functions (7c7e0e0)

  • Split moonwatcher into classification and detection object (d6e786c)

Testing

  • Added test for 0-dimensional scalar tensors (070f100)

  • Updated tests (0838969)

  • Updated tests for annotations.py (aae01fc)

v0.0.0 (2024-10-28)

Continuous Integration

  • Added automatic changelog workflow (8579ec7)

Documentation

  • Updated readme to include pre-commit hooks (24a2237)

Refactoring

  • Consolidate MoonwatcherDataset and MoonwatcherModel into unified Moonwatcher object (c231db6)