Teleprompter Base Class — Teleprompter • dsprrr

Base S7 class for optimization strategies (teleprompters). Teleprompters are responsible for optimizing modules by adjusting their prompts, demonstrations, or other parameters based on training data.

Usage

Teleprompter(metric = NULL, metric_threshold = NULL, max_errors = 5L)

Arguments

metric: Optional metric function for evaluating predictions. Subclasses that evaluate candidates document whether a metric is required.
metric_threshold: Minimum score required to be considered successful. If NULL, uses the metric's default threshold.
max_errors: Maximum number of errors allowed during optimization. Default is 5.

Value

A Teleprompter optimization strategy object.

Examples

Teleprompter()
#> <dsprrr::Teleprompter>
#>  @ metric          : NULL
#>  @ metric_threshold: NULL
#>  @ max_errors      : int 5