TotoForecaster

class TotoForecaster(seed=None, num_samples: int = 1, samples_per_batch: int = 1, prediction_type: str = 'median', scale_factor_exponent: int = 10, stabilize_with_global: bool = True, use_memory_efficient_attention: bool = False, model_path: str = 'Datadog/Toto-Open-Base-1.0', device=None)[source]

Toto foundation model forecaster for zero-shot forecasting.

Direct interface to forecaster from DataDog/toto [1].

Toto is a foundation model for multivariate time series forecasting with a focus on observability metrics. This model leverages innovative architectural designs to efficiently handle the high-dimensional, complex time series that are characteristic of observability data. Generate both point forecasts and uncertainty estimates using a Student-T mixture model. Support for variable prediction horizons and context lengths.

Known-future exogenous variables X are supported via Toto’s native exogenous mechanism: the columns of X are appended after the target channels as exogenous variates. When X is used, it must be supplied for every step of the forecast horizon, i.e. for all steps 1 .. max(fh) ahead of the cutoff (no gaps), since Toto consumes the known future values at each autoregressive step.

Parameters:

num_samplesint: Number of samples for probabilistic forecasting
samples_per_batchint, optional (default=1): Control memory usage during inference
prediction_typestring, optional (default=’median’): Type of prediction to generate (‘mean’ or ‘median’).
scale_factor_exponentint, optional (default=10): Exponent for the scale factor used in the model.
stabilize_with_globalboolean, optional (default=True): Whether to stabilize the model with global context.
use_memory_efficient_attentionboolean, optional (default=True): Whether to use memory-efficient attention mechanisms using Xformers.
model_pathstring, optional (default=’Datadog/Toto-Open-Base-1.0’): Path to the Toto huggingface model.
devicestring, optional (default=None): Specifies the device on which to run the model on (‘cpu’ or ‘cuda’).

Attributes:

cutoff: Cut-off = “present time” state of forecaster.
fh: Forecasting horizon that was passed.
is_fitted: Whether fit has been called.
state: State of the estimator.

References

[1]

https://github.com/DataDog/toto

Examples

>>> from sktime.datasets import load_longley
>>> from sktime.forecasting.toto import TotoForecaster
>>> _, y = load_longley()
>>> model = TotoForecaster()
>>> model.fit(y)
TotoForecaster()
>>> forecast = model.predict(fh=[1,2,5])

With known-future exogenous variables:

>>> from sktime.forecasting.model_selection import temporal_train_test_split
>>> X, y = load_longley()
>>> y_train, _, X_train, X_test = temporal_train_test_split(y, X, test_size=3)
>>> model = TotoForecaster()
>>> model.fit(y_train, X=X_train)
TotoForecaster()
>>> forecast = model.predict(fh=[1, 2, 3], X=X_test)

Methods

`check_is_fitted`([method_name])	Check if the estimator has been fitted.
`clone`()	Obtain a clone of the object with same hyper-parameters and config.
`clone_tags`(estimator[, tag_names])	Clone tags from another object as dynamic override.
`create_test_instance`([parameter_set])	Construct an instance of the class, using first test parameter set.
`create_test_instances_and_names`([parameter_set])	Create list of all test instances and a list of names for them.
`fit`(y[, X, fh])	Fit forecaster to training data.
`fit_predict`(y[, X, fh, X_pred])	Fit and forecast time series at future horizon.
`get_class_tag`(tag_name[, tag_value_default])	Get class tag value from class, with tag level inheritance from parents.
`get_class_tags`()	Get class tags from class, with tag level inheritance from parent classes.
`get_config`()	Get config flags for self.
`get_fitted_params`([deep])	Get fitted parameters.
`get_param_defaults`()	Get object's parameter defaults.
`get_param_names`([sort])	Get object's parameter names.
`get_params`([deep])	Get a dict of parameters values for this object.
`get_pretrained_params`([deep])	Get pretrained parameters of this estimator.
`get_tag`(tag_name[, tag_value_default, ...])	Get tag value from instance, with tag level inheritance and overrides.
`get_tags`()	Get tags from instance, with tag level inheritance and overrides.
`get_test_params`([parameter_set])	Return testing parameter settings for the estimator.
`is_composite`()	Check if the object is composed of other BaseObjects.
`load_from_path`(serial)	Load object from file location.
`load_from_serial`(serial)	Load object from serialized memory container.
`predict`([fh, X])	Forecast time series at future horizon.
`predict_interval`([fh, X, coverage])	Compute/return prediction interval forecasts.
`predict_proba`([fh, X, marginal])	Compute/return fully probabilistic forecasts.
`predict_quantiles`([fh, X, alpha])	Compute/return quantile forecasts.
`predict_residuals`([y, X])	Return residuals of time series forecasts.
`predict_var`([fh, X, cov])	Compute/return variance forecasts.
`pretrain`(y[, X, fh])	Pre-train forecaster on panel (global) data.
`reset`()	Reset the object to a clean post-init state.
`save`([path, serialization_format])	Save serialized self to bytes-like object or to (.zip) file.
`score`(y[, X, fh])	Scores forecast against ground truth, using MAPE (non-symmetric).
`set_config`(**config_dict)	Set config flags to given values.
`set_params`(**params)	Set the parameters of this object.
`set_random_state`([random_state, deep, ...])	Set random_state pseudo-random seed parameters for self.
`set_tags`(**tag_dict)	Set instance level tag overrides to given values.
`update`(y[, X, update_params])	Update cutoff value and, optionally, fitted parameters.
`update_predict`(y[, cv, X, update_params, ...])	Make predictions and update model iteratively over the test set.
`update_predict_single`([y, fh, X, update_params])	Update model with new data and make forecasts.