carate.runner package

Submodules

carate.runner.run module

class carate.runner.run.Run(dataset_name: str, num_features: int, num_classes: int, result_save_dir: str, model_save_freq: int, resume: bool, normalize: bool, data_set: DatasetObject, Evaluation: Evaluation, model_net: Model, optimizer: Optimizer, device: device, override: bool, logger: Any, net_dimension: int = 364, learning_rate: float = 0.0005, dataset_save_path: str = '.', test_ratio: int = 20, batch_size: int = 64, shuffle: bool = True, num_cv: int = 5, num_epoch: int = 150, num_heads: int = 3, dropout_forward: float = 0.6, dropout_gat: float = 0.5, custom_size: int | None = None)

Bases: DefaultObject

Run module to parametrize different tests and benchmarks from the command line

run() → None

Function to run training a model. Here only the CV is considered #TODO Make it more flexibile by passing the function as a parameter

class carate.runner.run.RunInitializer

Bases: object

classmethod from_file(config_filepath: str) → Run

classmethod from_json(json_object: Dict[Any, Any]) → Run

Module contents