p2pfl.node_state module¶

Node state.

class p2pfl.node_state.NodeState(addr)[source]¶

Bases: object

Class to store the main state of a learning node.

addr¶: The address of the node.

status¶: The status of the node.

learner¶: The learner of the node.

models_aggregated¶: The models aggregated by the node.

nei_status¶: The status of the neighbors.

train_set¶: The train set of the node.

train_set_votes¶: The votes of the train set.

train_set_votes_lock¶: The lock for the train set votes.

start_thread_lock¶: The lock for the start thread.

wait_votes_ready_lock¶: The lock for the wait votes ready.

model_initialized_lock¶: The lock for the model initialized.

Parameters:: addr (str) – The address of the node.

clear()[source]¶

Clear the state.

Return type:: None

property exp_name: str | None¶: Get the actual experiment name.

increase_round()[source]¶

Increase the round number.

Raises:: ValueError – If the experiment is not initialized.
Return type:: None

property round: int | None¶: Get the round.

set_experiment(exp_name, total_rounds, dataset_name=None, model_name=None, aggregator_name=None, framework_name=None, learning_rate=None, batch_size=None, epochs_per_round=None)[source]¶

Start a new experiment.

Parameters:

exp_name (str) – The name of the experiment.
total_rounds (int) – The total rounds of the experiment.
dataset_name (str | None) – The name of the dataset.
model_name (str | None) – The name of the model.
aggregator_name (str | None) – The name of the aggregator.
framework_name (str | None) – The name of the framework.
learning_rate (float | None) – The learning rate.
batch_size (int | None) – The batch size.
epochs_per_round (int | None) – The number of epochs per round.

Return type:

None

property total_rounds: int | None¶: Get the total rounds.