v1.19.0

Changes

• Migrate linting and formatting to ruff.
• Migrate test suite to idiomatic pytest (plain asserts, pytest.raises, capsys).

Breaking change

• Remove the deprecated positional experiment_name argument from MLFlowLogger. It must now be passed as a keyword argument.

v1.18.0

Changes

• Migrate to uv for dependency management and asdf for toolchain versioning.
• Consolidate all packaging metadata and dependency groups into a single pyproject.toml (hatchling build backend).
• Drop support for Python 3.8 and 3.9; add support for Python 3.13 and 3.14.
• Transfer repository ownership from GRAAL-Research to freud14.
• Fix SKLearnMetrics sample weight handling for scikit-learn >= 1.6.

v1.17.4

What's Changed

• Pickle module by @atremblay in #176

Full Changelog: v1.17.3...v1.17.4

v1.17.3

What's Changed

• Fix _TORCH_GREATER_EQUAL_1_12 with lightning_utilities by @atremblay in #174
• Add support for Python 3.12 by @freud14 in #175

Full Changelog: v1.17.2...v1.17.3

v1.17.2

What's Changed

• fix: np.Inf was deprecated forever and is now gone by @dhdaines in #171
• Update Black, isort, PyLint and flake8. by @freud14 in #172
• Update Github Actions by @freud14 in #173

New Contributors

• @dhdaines made their first contribution in #171

Full Changelog: v1.17.1...v1.17.2

v1.17.1

• Fix _XLA_AVAILABLE import with old versions of torchmetrics.
• Fix WandB tests.

v1.17

• FBeta is using the non-deterministic torch function bincount. Either by passing the argument make_deterministic to the FBeta class or by using one of the PyTorch functions torch.set_deterministic_debug_mode or torch.use_deterministic_algorithms, you can now make this function deterministic. Note that this might make your code slower.

v1.16

• Add run_id and terminate_on_end arguments to MLFlowLogger.

Breaking change:

• In MLFlowLogger, except for experiment_name, all arguments must now be passed as keyword arguments. Passing experiment_name as a positional argument is also deprecated and will be removed in future versions.

v1.15

• Remove support for Python 3.7

v1.14

• Update examples using classification metrics from torchmetrics to add the now required task argument.
• Fix the no LR scheduler bug when using PyTorch 2.0.