sgnlp.models.ufd.utils.load_trained_models

load_trained_models(cfg: sgnlp.models.ufd.data_class.UFDArguments, source_domain: str, target_language: str, target_domain: str) → Tuple[sgnlp.models.ufd.modeling.UFDAdaptorDomainModel, sgnlp.models.ufd.modeling.UFDAdaptorGlobalModel, sgnlp.models.ufd.modeling.UFDCombineFeaturesMapModel, sgnlp.models.ufd.modeling.UFDClassifierModel]

Helper function to load pretrained config and model weights for both supervised and unsupervised

models.

Parameters

• cfg (UFDArguments) – UFDArguments config load from configuration file

• source_domain (str) – name of source domain

• target_language (str) – name of target language

• target_domain (str) – name of target domain

Returns

return all supervised, unsupervised models with pretrained weights loaded.

Return type

Tuple[ UFDAdaptorDomainModel, UFDAdaptorGlobalModel, UFDCombineFeaturesMapModel, UFDClassifierModel, ]