nlptest.modelhandler.modelhandler.ModelFactory#
- class ModelFactory(model: str, task: str, hub: str, *args, **kwargs)#
Bases:
object
A factory class for instantiating models.
- __init__(model: str, task: str, hub: str, *args, **kwargs)#
Initializes the ModelFactory object. :param model: path to the model to evaluate :type model: str :param task: task to perform :type task: str
- Raises:
ValueError – If the task specified is not supported.
ValueError – If the given model is not supported.
Methods
__init__
(model, task, hub, *args, **kwargs)Initializes the ModelFactory object.
load_model
(task, hub, path, *args, **kwargs)Loads the model.
predict
(text, **kwargs)Perform predictions on input text.
predict_raw
(text)Perform predictions on input text.
Attributes
SUPPORTED_HUBS
SUPPORTED_MODULES
SUPPORTED_TASKS
- classmethod load_model(task: str, hub: str, path: str, *args, **kwargs) ModelFactory #
Loads the model.
- Parameters:
path (str) – path to model to use
task (str) – task to perform
hub (str) – model hub to load custom model from the path, either to hub or local disk.
Returns
- predict(text: str, **kwargs) NEROutput | SequenceClassificationOutput #
Perform predictions on input text.
- Parameters:
text (str) – Input text to perform predictions on.
- Returns:
predicted output
- Return type:
- predict_raw(text: str) List[str] #
Perform predictions on input text.
- Parameters:
text (str) – Input text to perform predictions on.
- Returns:
Predictions.
- Return type:
List[str]