v0.0.4

v0.0.4 (2025-01-02)

Bug Fixes

• Add checkpoints and checkpoint_load_func as arguments to signature of base classes (242a613, cca0c23)

• Fixing benchmarks (incl. some critical fixes) (818e5f0, 0963b0b, 1afb4ff, 1361a28, 466f03a, e746691, 1386da2, 4b9ee51)

• Randomize checkpoints properly in model randomization test (7df656d)

• Typing corrections (6bd80cb)

• Miscellaneous minor fixes (dac2c2d, c6bdb33, 12e2e06,7e18abf, 055695f, 1ec9d65, 0197506)

Chores

• Drop py3.8 support (5bfdfcf)

• Release workflow (b816e2e)

• Prevent major release bump in versions 0.y.x (e245bf7)

• Misc chores (d4135ec, 6591796,d345e43, 2c32b8f)

Code Style

• Introduce ruff (1ea6022, 0e329e5, 2af9e02)

• Ruff checks & formatting (cd4eab8, d2529d6, 708965b)

• Set line length to 79 (ce31af6)

Documentation

• Conventional Commits are now introduced - documented in CONTRIBUTING.md (db904f0)

• Typos (f827cfa, 2a669f7)

• Update README.md (f237f3d)

Testing

• Fixing tests (1a122ba, cc950ae, 11eed73)

---

Detailed Changes: v0.0.3...v0.0.4

v0.0.3

This release focuses on minor improvements, including updates to the documentation, enhancements to the README, and the addition of new workflows for better library management.

Please note that this release will not be published to PyPi, as there were no changes to user-facing functionality.

Stay tuned for upcoming releases that will introduce exciting new features 👀

v0.0.2

Our first release! 🚀

We’re thrilled to announce the first release of quanda, now available on PyPI.

What's included:

• Initial implementation of core functionality.
• A first suite of benchmarks with an MNIST model.
• Ready to install using pip install quanda.

Thank you for supporting the project!