config

class transfer_nlp.plugins.config.ExperimentConfig(experiment: Union[str, pathlib.Path, Dict], **env)

This file contains all necessary plugins classes that the framework will use to let a user interact with custom models, data loaders, etc…

The Registry pattern used here is inspired from this post: https://realpython.com/primer-on-python-decorators/

class transfer_nlp.plugins.config.ConfigFactoryABC

class transfer_nlp.plugins.config.ParamFactory(param)

Factory for simple parameters

class transfer_nlp.plugins.config.PluginFactory(cls, param2config_key: Optional[Dict[str, str]], *args, **kwargs)

Factory for complex objects creation

exception transfer_nlp.plugins.config.UnconfiguredItemsException(items)