golf_federated.client.process.config.model package

golf_federated.client.process.config.model.base module

class golf_federated.client.process.config.model.base.BaseModel(module: object, train_data: numpy.ndarray, train_label: numpy.ndarray, process_unit: str)

Bases: object

Model object class, the class function supports the main operation of model on Client.

choose_layer(prob_list: List) → List

Get the model parameter and set some layers to None based on the specified probability, i.e. some layers are not uploaded.

Args:

prob_list (list): Probability list, which corresponds to the parameter layers individually.

Returns:

List: Model parameters after adjustment.

get_weight() → List

Get model weight.

Returns:

List: Model weight.

abstract predict(data: numpy.ndarray) → numpy.ndarray

Abstract method for model prediction.

Args:

data (numpy.ndarray): Data values for prediction.

Returns:

Numpy.ndarray: Prediction result.

abstract train() → None

Abstract method for model training.

update_weight(new_weight: List) → None

Update model weight.

Args:

new_weight (List): Model weight for update.

golf_federated.client.process.config.model.tfmodel module

golf_federated.client.process.config.model.torchmodel module