Controls whether encodings should be generated in :func:`FutureCovariatesLocalForecastingModel.predict()``

The input and output target series window lengths consumed (input + output for training, input for prediction) or generated (output f

Load the electricity dataset as a list of univariate series, one for each household.

load the Uber TLC dataset as a list of univariate timeseries, one for each locationID.

Load the ExchangeRateDataset dataset as a list of univariate timeseries, one for each country.

Load the TrafficDataset dataset as a list of univariate timeseries, one for each ID.

Load the WeatherDataset dataset as a list of univariate timeseries, one for weather indicator.

Performs the exact same (unscaled) attention as the above function, but using the fast and fused F.scaled_dot_product_attention kernel.

Applies component masking to `ts_fit`.

Applies component masking to `ts_inverse_transform`.

Applies component masking to `ts_transform`.

Validator for boolean values.

Validator for positive integers.

Decorator allowing to specify some sanity check method(s) to be used on a class method. The decorator guarantees that args and kwargs from th

`CyclicTemporalEncoder` should not be transformed. Returns two elements for sine and cosine waves.

`DatetimeAttributeEncoder` accepts transformations

`IntegerIndexEncoder` accepts transformations. Note that transforming 'relative' `IntegerIndexEncoder` will return the absolute position (in t

`CallableIndexEncoder` accepts transformations.

Whether the `SingleEncoder` sub class accepts to be transformed.

Accuracy Score [1]_. For the true series :math:`y` and predicted series :math:`\\hat{y}` of length :math:`T`, it is computed per component/co

Absolute Error (AE). For the true series :math:`y` and predicted series :math:`\\hat{y}` of length :math:`T`, it is computed per component/co

Absolute Percentage Error (APE). For the true series :math:`y` and predicted series :math:`\\hat{y}` of length :math:`T`, it is computed as a

Absolute Ranged Relative Error (ARRE). For the true series :math:`y` and predicted series :math:`\\hat{y}` of length :math:`T`, it is computed as

Absolute Scaled Error (ASE) (see [1]_ for more information on scaled forecasting errors). It is the Absolute Error (AE) scaled by the Mean AE (MA

Area Under Tolerance Curve (AUTC). AUTC measures the overall alignment between actual and predicted series across a range of tolerance levels.

Returns the index generator base component name. - "pc": past covariates - "fc": future covariates

Returns the base encoder base component name. The string follows the given format: `"darts_enc_{covariates_temp}_{encoder}_{attribute}"`, wher

The bottom level component names of this series, or `None` if the series has no hierarchy.

The series containing the bottom-level components of this series in the same order as they appear in the series, or `None` if the series has n

List of categorical static variables in model

Raise :class:`optuna.TrialPruned` manually if pruned. Currently, ``intermediate_values`` are not properly propagated between processes due to

Returns the classes of the classifier model if the model was previously trained.

This decorator adds support for classification metrics including sanity checks and handling of class probabilities and categorical samples.

Coefficient of Variation (percentage). For the true series :math:`y` and predicted series :math:`\\hat{y}` of length :math:`T`, it is computed pe

The component (column) names of the series, as a ``pandas.Index``.

Applies component masking to the series fed to any `transform` method, and then reverts the masking for the final output series.

Generates names for the parameters of the Likelihood.

The component (column) names of the series, as a ``pandas.Index``.

Returns the covariates component names generated by `SequentialEncoder.past_encoders` and `SequentialEncoder.future_encoders`. A tuple of (pas

Returns the encoded component names. Only available after `Encoder.encode_train()` or `Encoder.encode_inference()` have been called.

We are re-defining a custom loss (which is not a likelihood loss) compared to Likelihood Parameters ---------- model

configures optimizers and learning rate schedulers for model optimization.

Confusion Matrix (CM) [1]_. For the true series :math:`y` and predicted series :math:`\\hat{y}` of length :math:`T`, it is computed per compo

Whether the model considers static covariates, if there are any.

Continuous Ranked Probability Score (CRPS). CRPS is a proper scoring rule that generalises the Mean Absolute Error (MAE) to probabilistic forecas