To listen to the message and handle them accordingly (used for \ distributed mode)

save the evaluation results of each client when the fl course \ early stopped or terminated

if file name not specified, use time stamp. Arguments: data (PyG.Data): the graph to be saved. p2root (str): the path of data

split given line/phrase into list of words Arguments: line: string representing phrase to be split :returns: list of strings,

Split dataset to train, val, test if not provided. Returns: List: List of split dataset, like ``[train, val, test]``

TOKENIZER = { 'shakespeare': { 'x': word_to_indices, 'y': letter_to_vec },

Interface for feature engineering, the default operation is none

This function is to perform log transform for data; Args: worker: ``federatedscope.core.workers.Worker`` to be wrapped Returns:

This function is to perform min-max scale vfl tabular data. Args: worker: ``federatedscope.core.workers.Worker`` to be wrapped R

This function is to perform quantile_binning for vfl tabular data. Args: worker: ``federatedscope.core.workers.Worker`` to be wrapped

This function is to perform z-norm/standardization for vfl tabular data. Args: worker: ``federatedscope.core.workers.Worker`` to be w

This function is to perform uniform_binning for vfl tabular data. Args: worker: ``federatedscope.core.workers.Worker`` to be wrapped

This function is to perform z-norm/standardization for vfl tabular data. Args: worker: ``federatedscope.core.workers.Worker`` to be w

r""" generate Cifar10 Dataset transform and split dict for contrastive learning return { 'client_id': { 't

r""" generate Cifar10 Dataset transform and split dict for linear prob evaluation of contrastive learning return { 'client

r"""Mend the graph with generation model Arguments: model (torch.nn.module): trained generation model impaired_data (PyG.Data): im

Remove enclosing quotation marks, if any.

Generate getter and setter for a directory property.

ROUGE can evaluate several system summaries for a given text against several model summaries, i.e. there is an m-to-n relatio

Get supplied command line arguments for ROUGE or use default ones.

Arguments: cfg (federatedscope.core.configs.config.CN): dict \ like object, where each key-value pair cor

r""" FedSage+ consists of three of training stages. Stage1: 0, local pre-train for generator. Stage2: -> 2 * fedgen_epoch, fed

Automatically generate the properties for directories.

Verfify presence of ROUGE-1.5.5.pl and data folder, and set those paths.

Apply MultiKrum to select \theta (\theta <= client_num- 2*self.byzantine_node_num) local models