
############################## Experiment details ##############################

dataset: femnist
dataset_dir: ../data/femnist/
output_dir: output/fedavg_cnn-2022-08-03_22:53:38/
model: cnn
num_rounds: 100
eval_every: 10
ServerType: <class 'fedavg_cnn_exp.fedavg_cnn.<locals>.make_fedavg_server_client.<locals>.FedAvgServer'>
client_types: [(<class 'fedavg_cnn_exp.fedavg_cnn.<locals>.make_fedavg_server_client.<locals>.FedAvgClient'>, 100)]
clients_per_round: 10
client_lr: 0.06
batch_size: 64
seed: 0
use_val_set: False
num_epochs: 10
gpus_per_client_cluster: 1
num_client_clusters: 2
save_model: False

############################## Simulation ##############################

Spawning 2 ClientClusters using cuda device (this may take a while)
100 total clients: 100 FedAvgClients

>>> Training Accuracy: Round 0 <<<
+------------+------------+-------------------+-------------------+
|    average |     median |   10th percentile |   90th percentile |
|------------+------------+-------------------+-------------------|
| 0.00254635 | 0.00292398 |                 0 |        0.00377645 |
+------------+------------+-------------------+-------------------+
>>> Test Accuracy: Round 0 <<<
+------------+----------+-------------------+-------------------+
|    average |   median |   10th percentile |   90th percentile |
|------------+----------+-------------------+-------------------|
| 0.00330761 |        0 |                 0 |         0.0239286 |
+------------+----------+-------------------+-------------------+
--- Round 1 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 2 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 3 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 4 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 5 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 6 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 7 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 8 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 9 of 100: Training 10 clients: 10 FedAvgClients ---
--- Round 10 of 100: Training 10 clients: 10 FedAvgClients ---
