roxanneluo / federated-learning

Federated-Learning

We implement the paper Communication-Efficient Learning of Deep Networks from Decentralized Data. Its blog is here.

We wish to explore several possible approaches:

1. server pulling gradient update from clients;
2. clients eagerly train and push updates to server;
3. extend to multi-server; servers gossip to sync;
4. client-only p2p gradient updates

federated-learning's Issues

Issue while running

Hi ,
I keep getting this error:
Traceback (most recent call last):
File "fl_client.py", line 214, in
FederatedClient("127.0.0.1", 5000, datasource.Mnist)
File "fl_client.py", line 92, in init
self.sio.wait()
File "C:\Users\zxcve\Anaconda3\lib\site-packages\socketIO_client_init_.py", line 251, in wait
self.process_packets()
File "C:\Users\zxcve\Anaconda3\lib\site-packages\socketIO_client_init.py", line 278, in _process_packets
self.process_packet(engineIO_packet)
File "C:\Users\zxcve\Anaconda3\lib\site-packages\socketIO_client_init.py", line 491, in process_packet
delegate(parse_socketIO_packet_data(socketIO_packet_data), namespace)
File "C:\Users\zxcve\Anaconda3\lib\site-packages\socketIO_client_init.py", line 513, in _on_event
namespace._find_packet_callback(event)(*args)
File "fl_client.py", line 139, in on_request_update
my_weights, train_loss, train_accuracy = self.local_model.train_one_round()
File "fl_client.py", line 57, in train_one_round
validation_data=(self.x_valid, self.y_valid))
File "C:\Users\zxcve\Anaconda3\lib\site-packages\keras\models.py", line 893, in fit
initial_epoch=initial_epoch)
File "C:\Users\zxcve\Anaconda3\lib\site-packages\keras\engine\training.py", line 1555, in fit
batch_size=batch_size)
File "C:\Users\zxcve\Anaconda3\lib\site-packages\keras\engine\training.py", line 1413, in _standardize_user_data
exception_prefix='target')
File "C:\Users\zxcve\Anaconda3\lib\site-packages\keras\engine\training.py", line 142, in _standardize_input_data
str(array.shape))
ValueError: Error when checking target: expected dense_2 to have 2 dimensions, but got array with shape (1200, 1, 10)
disconnect

Do you know fix for this?

could you introduce how to use this REPO ?

if you could ,i hope you can write down in the README.

I encounter some version control problems

Sorry to bother you, could you please tell me the version of the following package: socketio-client, flask, flask-socketio.

could you please introduce how to use this REPO ?