nemo_rl.data.datasets.preference_datasets.tulu3#

Module Contents#

Classes#

Tulu3PreferenceDataset

Tulu3 preference dataset for DPO training.

Functions#

API#

nemo_rl.data.datasets.preference_datasets.tulu3.to_preference_data_format(
data: dict[str, Any],
) dict[str, list[dict[str, int | list[dict[str, str | Any]]]] | list[dict[str, str]]]#
class nemo_rl.data.datasets.preference_datasets.tulu3.Tulu3PreferenceDataset#

Tulu3 preference dataset for DPO training.

Initialization