graphorge.gnn_base_model.train.training.load_training_state

load_training_state(model, opt_algorithm, optimizer, load_model_state=None, is_remove_posterior=True)

Load model and optimizer states from available training data.

Material patch model state file is stored in model_directory under the name < model_name >.pt, < model_name >-< epoch >.pt, < model_name >-best.pt or < model_name >-< epoch >-best.pt.

Optimizer state file is stored in model_directory under the name < model_name >_optim.pt or < model_name >_optim-< epoch >.pt.

Both model and optimizer are updated ‘in-place’ with loaded state data.

Parameters:

• model (torch.nn.Module) – Model.

• opt_algorithm ({'adam',}, default='adam') –

  Optimization algorithm:

  ’adam’ : Adam (torch.optim.Adam)

• optimizer (torch.optim.Optimizer) – PyTorch optimizer.

• load_model_state ({'best', 'last', int, None}, default=None) –

  Load available Graph Neural Network model state from the model directory. Options:

  ’best’ : Model state corresponding to best performance available

  ’last’ : Model state corresponding to highest training epoch

  int : Model state corresponding to given training epoch

  None : Model default state file

• is_remove_posterior (bool, default=True) – Remove material patch model state files corresponding to training epochs posterior to the loaded state file. Effective only if loaded training epoch is known.

Returns:

loaded_epoch – Training epoch corresponding to loaded state data. Defaults to 0 if training epoch is unknown.

Return type:

int