Validation
- cloudcasting.validation.validate(model: AbstractModel, data_path: list[str] | str, wandb_project_name: str, wandb_run_name: str, nan_to_num: bool = False, batch_size: int = 1, num_workers: int = 0, batch_limit: int | None = None) None
Run the full validation procedure on the model and log the results to wandb.
- Parameters:
model (AbstractModel) – the model to be validated
data_path (str) – path to the validation data set
nan_to_num (bool, optional) – Whether to convert NaNs to -1. Defaults to False.
batch_size (int, optional) – Defaults to 1.
num_workers (int, optional) – Defaults to 0.
batch_limit (int | None, optional) – Defaults to None. For testing purposes only.
- cloudcasting.validation.validate_from_config(config_file: ~typing.Annotated[str, <typer.models.OptionInfo object at 0x7f349c697790>] = 'validate_config.yml', model_file: ~typing.Annotated[str, <typer.models.OptionInfo object at 0x7f349e2cf110>] = 'model.py') None
CLI function to validate a model from a config file. Example templates of these files can be found at https://github.com/alan-turing-institute/ocf-model-template.