v1.1.0

Release 1.1.0

This release adds the packaged V25 Auto Router as the default Tabnetics entry point. The router model is bundled inside the Python package, uses descriptors computed from the training data only, and selects among supported validation-backed profiles without any network download.

Install from PyPI:

pip install tabnetics==1.1.0

Highlights:

• DFFSConfig(auto_router_enabled=True) is now the default path.
• The V25 calibrated score-router predicts balanced accuracy and macro-F1 for supported candidate profiles, then applies a conservative calibrated policy.
• Public documentation now includes the Auto Router model card, rationale, and usage examples.
• The interactive results browser includes V25 training-set router evidence and the available frozen-router holdout context.

Documentation:

• https://tabnetics.org/
• https://tabnetics.org/AUTO_ROUTER.html
• https://tabnetics.org/results-browser.html
• https://tabnetics.org/RESULTS.html