Inference tuning

`TuningConfig`

Configuration for tuning the model during fit/predict calls.

TuningConfig(
    calibrate_temperature: bool = False,
    tuning_holdout_frac: Literal["auto"] | float = "auto",
    tuning_n_folds: Literal["auto"] | int = "auto",
)

Fields

Name	Type	Default	Description
`calibrate_temperature`	`bool`	`False`	Whether to calibrate the softmax temperature. Set to `True` to enable.
`tuning_holdout_frac`	`Literal[“auto”] \| float`	`"auto"`	The percentage of the data to hold out for tuning per split. If “auto”, a value is automatically chosen based on the dataset size, trading off between computational cost and accuracy.
`tuning_n_folds`	`Literal[“auto”] \| int`	`"auto"`	The number of cross-validation folds to use for tuning. If “auto”, a value is automatically chosen based on the dataset size, trading off between computational cost and accuracy.

`TuningConfig.resolve`

View source

Resolves ‘auto’ values based on the number of samples.

TuningConfig.resolve(
    num_samples: int,
) -> Self

Parameters

Parameter	Type	Default	Description
`num_samples`	`int`	Required	The number of samples in the training data.

Returns

Type	Description
`Self`	A new `TuningConfig` instance with resolved values.

`ClassifierTuningConfig`

View source

Configuration for tuning the model during fit/predict calls for classification tasks.

ClassifierTuningConfig(
    calibrate_temperature: bool = False,
    tuning_holdout_frac: Literal["auto"] | float = "auto",
    tuning_n_folds: Literal["auto"] | int = "auto",
    tune_decision_thresholds: bool = False,
)

Fields

Name	Type	Default	Description
`calibrate_temperature`	`bool`	`False`	Whether to calibrate the softmax temperature. Set to `True` to enable.
`tuning_holdout_frac`	`Literal[“auto”] \| float`	`"auto"`	The percentage of the data to hold out for tuning per split. If “auto”, a value is automatically chosen based on the dataset size, trading off between computational cost and accuracy.
`tuning_n_folds`	`Literal[“auto”] \| int`	`"auto"`	The number of cross-validation folds to use for tuning. If “auto”, a value is automatically chosen based on the dataset size, trading off between computational cost and accuracy.
`tune_decision_thresholds`	`bool`	`False`	Whether to tune decision thresholds for the specified `eval_metric`. Set to `True` to enable.

`ClassifierTuningConfig.resolve`

View source

Resolves ‘auto’ values based on the number of samples.

ClassifierTuningConfig.resolve(
    num_samples: int,
) -> Self

Parameters

Parameter	Type	Default	Description
`num_samples`	`int`	Required	The number of samples in the training data.

Returns

Type	Description
`Self`	A new `TuningConfig` instance with resolved values.

`RegressorTuningConfig`

View source

Configuration for tuning the model during fit/predict calls for regression tasks. This currently adds no fields beyond the ones inherited from [TuningConfig][tabpfn.inference_tuning.TuningConfig]. It exists so that the task type can be recovered from the configuration alone, and so that regression-only tuning options have a place to live as they are added.

RegressorTuningConfig(
    calibrate_temperature: bool = False,
    tuning_holdout_frac: Literal["auto"] | float = "auto",
    tuning_n_folds: Literal["auto"] | int = "auto",
)

Fields

Name	Type	Default	Description
`calibrate_temperature`	`bool`	`False`	Whether to calibrate the softmax temperature. Set to `True` to enable.
`tuning_holdout_frac`	`Literal[“auto”] \| float`	`"auto"`	The percentage of the data to hold out for tuning per split. If “auto”, a value is automatically chosen based on the dataset size, trading off between computational cost and accuracy.
`tuning_n_folds`	`Literal[“auto”] \| int`	`"auto"`	The number of cross-validation folds to use for tuning. If “auto”, a value is automatically chosen based on the dataset size, trading off between computational cost and accuracy.

`RegressorTuningConfig.resolve`

View source

Resolves ‘auto’ values based on the number of samples.

RegressorTuningConfig.resolve(
    num_samples: int,
) -> Self

Parameters

Parameter	Type	Default	Description
`num_samples`	`int`	Required	The number of samples in the training data.

Returns

Type	Description
`Self`	A new `TuningConfig` instance with resolved values.

`ClassifierEvalMetrics`

View source

Metric by which predictions will be ultimately evaluated on test data. Fields

Name	Type	Default	Description
`F1`	—	`"f1"`	—
`ACCURACY`	—	`"accuracy"`	—
`BALANCED_ACCURACY`	—	`"balanced_accuracy"`	—
`ROC_AUC`	—	`"roc_auc"`	—
`LOG_LOSS`	—	`"log_loss"`	—

`RegressorEvalMetrics`

View source

Metric by which predictions will be ultimately evaluated on test data. Fields

Name	Type	Default	Description
`NLL`	—	`"nll"`	Negative log-likelihood of the targets under the predicted bar distribution, evaluated in the raw target space.
`CRPS`	—	`"crps"`	Continuous ranked probability score of the targets under the predicted bar distribution, evaluated in the raw target space.

Reference

tabpfn

tabpfn-client

tabpfn-extensions

REST API

`TuningConfig`

`TuningConfig.resolve`

`ClassifierTuningConfig`

`ClassifierTuningConfig.resolve`

`RegressorTuningConfig`

`RegressorTuningConfig.resolve`

`ClassifierEvalMetrics`

`RegressorEvalMetrics`

​TuningConfig

​TuningConfig.resolve

​ClassifierTuningConfig

​ClassifierTuningConfig.resolve

​RegressorTuningConfig

​RegressorTuningConfig.resolve

​ClassifierEvalMetrics

​RegressorEvalMetrics

`TuningConfig`

`TuningConfig.resolve`

`ClassifierTuningConfig`

`ClassifierTuningConfig.resolve`

`RegressorTuningConfig`

`RegressorTuningConfig.resolve`

`ClassifierEvalMetrics`

`RegressorEvalMetrics`