Skip to content

Evaluation

🚧 Not yet implemented

This workflow is scaffolded but not yet implemented — the entry point (scripts/evaluate.py) is a stub that raises NotImplementedError. This page will fill in once evaluation lands. Until then, treat the config below as the intended shape.

scripts/evaluate.py will evaluate trained models on the Slovenian benchmarks (NER, sentiment, SuperGLUE, etc.) registered in the task config.

Entry point: scripts/evaluate.py Config: configs/data/tasks.yaml