Training¶

🚧 Not yet implemented

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

scripts/train.py will pretrain or fine-tune small language models from YAML configs under configs/training/, with MLflow experiment tracking.

Entry point: scripts/train.py Config: configs/training/pretrain.yaml (and configs/training/finetune_ner.yaml)