def __init__(self, ):
"""
Args:
transfer_variable:
converge_type: see federatedml/optim/convergence.py
tolerate_val:
"""
self.loss_scatter = loss_scatter.Server()
self.has_converged = has_converged.Server()
def __init__(self, trans_var):
super().__init__(trans_var=trans_var)
self.model = None
self.aggregator = secure_mean_aggregator.Server(self.transfer_variable.secure_aggregator_trans_var)
self.loss_scatter = loss_scatter.Server(self.transfer_variable.loss_scatter_trans_var)
self.has_converged = has_converged.Server(self.transfer_variable.has_converged_trans_var)
self._summary = dict(loss_history=[], is_converged=False)
def loss_scatter_call(job_id, role, ind, *args):
losses = args[0]
if role == consts.ARBITER:
losses = loss_scatter.Server().get_losses()
return list(losses)
elif role == consts.HOST:
loss = losses[ind + 1]
return loss_scatter.Client().send_loss(loss)
else:
loss = losses[0]
return loss_scatter.Client().send_loss(loss)
def __init__(self, trans_var=LegacyAggregatorTransVar()):
self._guest_parties = trans_var.get_parties(roles=[consts.GUEST])
self._host_parties = trans_var.get_parties(roles=[consts.HOST])
self._client_parties = trans_var.client_parties
self._loss_sync = loss_scatter.Server(trans_var.loss_scatter)
self._converge_sync = has_converged.Server(trans_var.has_converged)
self._model_scatter = model_scatter.Server(trans_var.model_scatter)
self._model_broadcaster = model_broadcaster.Server(
trans_var.model_broadcaster)
self._random_padding_cipher = random_padding_cipher.Server(
trans_var.random_padding_cipher)
def server_init_model(self, param):
self.aggregate_iteration_num = 0
self.aggregator = secure_mean_aggregator.Server(
self.transfer_variable.secure_aggregator_trans_var)
self.loss_scatter = loss_scatter.Server(
self.transfer_variable.loss_scatter_trans_var)
self.has_converged = has_converged.Server(
self.transfer_variable.has_converged_trans_var)
self._summary = dict(loss_history=[], is_converged=False)
self.param = param
self.enable_secure_aggregate = param.secure_aggregate
self.max_aggregate_iteration_num = param.max_iter
early_stop = self.model_param.early_stop
self.converge_func = converge_func_factory(early_stop.converge_func,
early_stop.eps).is_converge
self.loss_consumed = early_stop.converge_func != "weight_diff"
def __init__(self, verbose=False):
self.aggregator = secure_sum_aggregator.Server(
enable_secure_aggregate=True)
self.scatter = loss_scatter.Server()
self.verbose = verbose