feat: add OneIG benchmark with subset support#508

Open

davidberenstein1957 wants to merge 10 commits intomainfrom

feat/add-oneig-benchmark

Commits on Jan 22, 2026

feat: add benchmark support to PrunaDataModule and implement PartiPrompts benchmark
davidberenstein1957
committed

Commits on Jan 31, 2026

refactor: simplify benchmark system, extend PartiPrompts with subset filtering

davidberenstein1957
and
cursoragent
committed
fix: add Numpydoc parameter docs for BenchmarkInfo

davidberenstein1957
and
cursoragent
committed
feat: add benchmark discovery functions and expand benchmark registry

davidberenstein1957
and
cursoragent
committed
fix: use correct metric names from MetricRegistry

davidberenstein1957
and
cursoragent
committed
feat: add OneIG Text Rendering benchmark

davidberenstein1957
and
cursoragent
committed
feat: add OneIG Alignment benchmark

davidberenstein1957
and
cursoragent
committed
feat: add unified OneIG benchmark with subset support

davidberenstein1957
and
cursoragent
committed
fix: address review feedback for OneIG benchmark

davidberenstein1957
and
cursoragent
committed
fix: add empty dataset guard for OneIG subsets

davidberenstein1957
and
cursoragent
committed