NHITSHyperparams

nhits_model.NHITSHyperparams · inherits PydanticBaseModel

Tunable training knobs; aligned with TFT / shorter training.

Attributes

Attribute	Type	Default	Description
input_size	int	128	NeuralForecast input_size (lookback).
max_steps	int	25	Lightning optimization steps per fit (not forecast horizon; tasks cap h at 128).
batch_size	int	64	Minibatch size (series per step); larger improves GPU/CPU throughput until memory-bound.

NHITSModel

nhits_model.NHITSModel · inherits BaseModel

__init__(self, params, settings)

Parameter	Type	Default	Description
params	Dict[str, Any]	-	(undocumented)
settings	Dict[str, Any]	-	(undocumented)

train(self, y_context, y_target, timestamps_context, timestamps_target, x_context=None, x_target=None, **kwargs) -> NHITSModel

Parameters:

Parameter	Type	Default	Description
y_context	np.ndarray	-	(undocumented)
y_target	np.ndarray	-	(undocumented)
timestamps_context	np.ndarray	-	(undocumented)
timestamps_target	np.ndarray	-	(undocumented)
x_context	Optional[np.ndarray]	None	(undocumented)
x_target	Optional[np.ndarray]	None	(undocumented)
**kwargs	Any	-	(undocumented)

Returns: NHITSModel (the fitted model; sets self.is_fitted = True).

predict(self, y_context, timestamps_context, timestamps_target, x_context=None, x_target=None, **kwargs) -> np.ndarray

Parameters:

Parameter	Type	Default	Description
y_context	np.ndarray	-	(undocumented)
timestamps_context	np.ndarray	-	(undocumented)
timestamps_target	np.ndarray	-	(undocumented)
x_context	Optional[np.ndarray]	None	(undocumented)
x_target	Optional[np.ndarray]	None	(undocumented)
**kwargs	Any	-	(undocumented)

Returns: np.ndarray (The predicted forecast values).

Raises:

Exception	Description
ValueError	If train() has not been called prior to prediction.