rollout_metrics_noisy_ics

rollout_metrics_noisy_ics#

Attributes#

logger

description

Functions#

compute_metrics(metrics, y_pred, y, date_time, ...)

Compute metrics and update metrics_results.

calculate_ensemble_metrics(ensemble_preds, true_values)

Calculate RMSE and STD per channel

predict(rank, world_size, conf[, backend, p])

Module Contents#

rollout_metrics_noisy_ics.logger#
rollout_metrics_noisy_ics.compute_metrics(metrics, y_pred, y, date_time, forecast_step, utc_datetime)#

Compute metrics and update metrics_results.

rollout_metrics_noisy_ics.calculate_ensemble_metrics(ensemble_preds, true_values)#

Calculate RMSE and STD per channel

Parameters:

  • ensemble_preds – (n_members, batch, channels, 1, height, width)

  • true_values – (1, batch, channels, 1, height, width)

rollout_metrics_noisy_ics.predict(rank, world_size, conf, backend=None, p=None)#
rollout_metrics_noisy_ics.description = 'Rollout AI-NWP forecasts'#
Contents