Load Emulator Models

Load the neural network models into memory. Returns an EmulatorModels object that can be passed to generate_scenario_predictions() for lower-level control, or models are cached automatically for use by run_malaria_emulator().

Usage

load_emulator_models(
  predictor = "prevalence",
  device = NULL,
  verbose = TRUE,
  force_reload = FALSE
)

Arguments

predictor

Character. Which models to load: "prevalence" or "cases".

device

Character. Device to load models on ("cpu" or "cuda"). If NULL, auto-detected.

verbose

Logical. Print loading messages. Default TRUE.

force_reload

Logical. Force reload even if cached. Default FALSE.

Value

An EmulatorModels object containing loaded models and configuration. This can be passed to generate_scenario_predictions().

See also

generate_scenario_predictions(), run_malaria_emulator()

Examples

if (FALSE) { # \dontrun{
# Load prevalence models
models <- load_emulator_models(predictor = "prevalence")

# Use with generate_scenario_predictions
scenarios <- create_scenarios(...)
predictions <- generate_scenario_predictions(scenarios, models)
} # }