irtk.model.MultiOpt¶
- class irtk.model.MultiOpt(scene: Any, model_classes: List[type])¶
A model that combines multiple other models.
- _models¶
A list of model instances.
- __init__(scene: Any, model_classes: List[type]) None¶
Initialize the MultiOpt model.
- Parameters:
scene – The scene object associated with the model.
model_classes – A list of model classes to instantiate.
- get_regularization() torch.Tensor¶
Get the combined regularization from all models.
- Returns:
A torch tensor containing the combined regularization value.
- get_results() List[Any]¶
Get the results from all models.
- Returns:
A list of results from all models.
- initialize() None¶
Initialize all the models.
- load_states(state_path: str) None¶
Load the states for all models.
- Parameters:
state_path – The path to load the states from.
- save_states(state_path: str) None¶
Save the states for all models.
- Parameters:
state_path – The path to save the states to.
- schedule_lr(curr_iter: int) None¶
Schedule the learning rate for all models.
- Parameters:
curr_iter – The current iteration number.
- set_data() None¶
Set the data for all models.
- step() None¶
Perform a step for all models.
- write_results(result_path: str) None¶
Write the results from all models.
- Parameters:
result_path – The path to write the results to.
- zero_grad() None¶
Zero out the gradients for all models.