core.finetuning

How much of a pretrained encoder a downstream run is allowed to move.

FinetuningStrategy	Base class for finetuning strategies.
FullFinetuning	No-op strategy that leaves all model parameters trainable.
GradualUnfreezing	Freeze all parameters except those matching unfrozen_prefixes, then unfreeze at a target epoch.
Probe	Freeze all parameters except those matching unfrozen_prefixes.