credit.trainers.trainerERA5gen1

Attributes

logger
Trainer

Classes

TrainerERA5Gen1

Helper class that provides a standard way to create an ABC using

Module Contents

credit.trainers.trainerERA5gen1.logger

class credit.trainers.trainerERA5gen1.TrainerERA5Gen1(model: torch.nn.Module, rank: int, conf: dict)

Bases: credit.trainers.base_trainer.BaseTrainer

Helper class that provides a standard way to create an ABC using inheritance.

flag_mass_conserve = False

flag_water_conserve = False

flag_energy_conserve = False

opt_mass = None

opt_water = None

opt_energy = None

varnum_diag

static_dim_size

retain_graph

forecast_len

valid_history_len

valid_forecast_len

train_one_epoch(epoch, trainloader, optimizer, criterion, scaler, scheduler, metrics)

Train for one epoch.

Parameters:

• epoch – Current epoch number.

• conf – Full configuration dict (data keys accessed here for schema stability).

• trainloader – DataLoader for training.

• optimizer – Standard training objects.

• criterion – Standard training objects.

• scaler – Standard training objects.

• scheduler – Standard training objects.

• metrics – Standard training objects.

Returns:

Training metrics for the epoch.

Return type:

dict

validate(epoch, valid_loader, criterion, metrics)

Validate for one epoch.

Parameters:

• epoch – Current epoch number.

• conf – Full configuration dict.

• valid_loader – DataLoader for validation.

• criterion – Loss and metric callables.

• metrics – Loss and metric callables.

Returns:

Validation metrics for the epoch.

Return type:

dict

credit.trainers.trainerERA5gen1.Trainer