urbanlc.model.deep_learning.MSSDeepLearning.train_one_epoch#

MSSDeepLearning.train_one_epoch(step: int, train_loader: DataLoader, epoch: int, MAX_EPOCH: int, metrics: Dict[str, Any], GRADIENT_ACCUMULATION_FACTOR: Optional[int]) → Tuple[int, Dict[str, Any], float, int]#

Train the model for one epoch.

Parameters:

  • step (int) – Current training step.

  • train_loader (DataLoader) – DataLoader for training.

  • epoch (int) – Current training epoch.

  • MAX_EPOCH (int) – Maximum number of training epochs.

  • metrics (Dict[str, Any]) – Dictionary of metrics.

  • GRADIENT_ACCUMULATION_FACTOR (Optional[int]) – Gradient accumulation factor.

Returns:

Updated step, metrics, training loss, and sample count.

Return type:

Tuple[int, Dict[str, Any], float, int]