Add mistnet example and fix some bugs when running mistnet

Signed-off-by: XinYao1994 <xyao@cs.hku.hk>

examples/federated-learning-mistnet-client.Dockerfile

@@ -0,0 +1,18 @@
+FROM tensorflow/tensorflow:1.15.4
+
+RUN apt update \
+  && apt install -y libgl1-mesa-glx
+
+COPY ./lib/requirements.txt /home
+
+RUN pip install -r /home/requirements.txt
+
+ENV PYTHONPATH "/home/lib:/home/plato"
+
+COPY ./lib /home/lib
+COPY ./plato /home/plato
+
+WORKDIR /home/work
+COPY examples/federated_learning/mistnet/  /home/work/
+
+ENTRYPOINT ["python", "train_worker.py"]

examples/federated-learning-mistnet.Dockerfile

@@ -0,0 +1,18 @@
+FROM tensorflow/tensorflow:1.15.4
+
+RUN apt update \
+  && apt install -y libgl1-mesa-glx
+
+COPY ./lib/requirements.txt /home
+
+RUN pip install -r /home/requirements.txt
+
+ENV PYTHONPATH "/home/lib:/home/plato"
+
+COPY ./lib /home/lib
+COPY ./plato /home/plato
+
+WORKDIR /home/work
+COPY examples/federated_learning/mistnet/  /home/work/
+
+CMD ["/bin/sh", "-c", "ulimit -n 50000; python agg_worker.py"]

examples/federated_learning/mistnet/agg_worker.py

@@ -0,0 +1,8 @@
+import sedna.service.server
+from sedna.service.server import MistnetServer
+import os
+if __name__ == '__main__':
+    # run_server()
+    server = MistnetServer()
+    server.run()
+

examples/federated_learning/mistnet/client.yml

@@ -0,0 +1,68 @@
+clients:
+    # Type
+    type: mistnet
+
+    # The total number of clients
+    total_clients: 1
+
+    # The number of clients selected in each round
+    per_round: 1
+
+    # Should the clients compute test accuracy locally?
+    do_test: false
+
+server:
+    type: mistnet
+
+    address: 127.0.0.1
+    port: 7363
+
+data:
+    # The training and testing dataset
+    datasource: MNIST
+
+    # Where the dataset is located
+    data_path: ./data
+
+    # Number of samples in each partition
+    partition_size: 20000
+
+    # Fixed random seed
+    random_seed: 1
+
+    # IID, biased, or sharded?
+    sampler: iid
+
+trainer:
+    # The type of the trainer
+    type: basic
+
+    # The maximum number of training rounds
+    rounds: 1
+
+    # Whether the training should use multiple GPUs if available
+    parallelized: false
+
+    # The maximum number of clients running concurrently
+    max_concurrency: 3
+
+    # The target accuracy
+    target_accuracy: 0.95
+
+    # Number of epoches for local training in each communication round
+    epochs: 10
+    batch_size: 32
+    optimizer: SGD
+    learning_rate: 0.01
+    momentum: 0.9
+    weight_decay: 0.0
+
+    # The machine learning model
+    model_name: lenet5
+
+algorithm:
+    # Aggregation algorithm
+    type: mistnet
+
+    cut_layer: relu3
+    epsilon: null

examples/federated_learning/mistnet/server.yml

@@ -0,0 +1,68 @@
+clients:
+    # Type
+    type: mistnet
+
+    # The total number of clients
+    total_clients: 0
+
+    # The number of clients selected in each round
+    per_round: 1
+
+    # Should the clients compute test accuracy locally?
+    do_test: false
+
+server:
+    type: mistnet
+
+    address: 127.0.0.1
+    port: 7363
+
+data:
+    # The training and testing dataset
+    datasource: MNIST
+
+    # Where the dataset is located
+    data_path: ./data
+
+    # Number of samples in each partition
+    partition_size: 20000
+
+    # Fixed random seed
+    random_seed: 1
+
+    # IID, biased, or sharded?
+    sampler: iid
+
+trainer:
+    # The type of the trainer
+    type: basic
+
+    # The maximum number of training rounds
+    rounds: 1
+
+    # Whether the training should use multiple GPUs if available
+    parallelized: false
+
+    # The maximum number of clients running concurrently
+    max_concurrency: 3
+
+    # The target accuracy
+    target_accuracy: 0.95
+
+    # Number of epoches for local training in each communication round
+    epochs: 10
+    batch_size: 32
+    optimizer: SGD
+    learning_rate: 0.01
+    momentum: 0.9
+    weight_decay: 0.0
+
+    # The machine learning model
+    model_name: lenet5
+
+algorithm:
+    # Aggregation algorithm
+    type: mistnet
+
+    cut_layer: relu3
+    epsilon: null

examples/federated_learning/mistnet/train_worker.py

@@ -0,0 +1,9 @@
+import sedna.core.federated_learning
+from sedna.core.federated_learning import MistWorker
+import asyncio
+
+if __name__ == '__main__':
+    client = MistWorker()
+    client.configure()
+    loop = asyncio.get_event_loop()
+    loop.run_until_complete(client.start_client())

examples/federated_learning/surface_defect_detection/aggregation_worker/aggregate.py

@@ -30,12 +30,13 @@ def run_server():
     )
     server.start()
 
-
+import sedna.service.server
 from sedna.service.server import PlatoServer
 from torch import nn
 
 if __name__ == '__main__':
     # run_server()
+
     model = nn.Sequential(
         nn.Linear(28 * 28, 128),
         nn.ReLU(),

examples/federated_learning/surface_defect_detection/aggregation_worker/config.yml

@@ -0,0 +1,63 @@
+clients:
+    # Type
+    type: simple
+
+    # The total number of clients, server
+    total_clients: 0
+
+    # The number of clients selected in each round
+    per_round: 1
+
+    # Should the clients compute test accuracy locally?
+    do_test: true
+
+server:
+    address: 127.0.0.1
+    port: 7363
+
+data:
+    # The training and testing dataset
+    datasource: MNIST
+
+    # Where the dataset is located
+    data_path: ./data
+
+    # Number of samples in each partition
+    partition_size: 20000
+
+    # IID or non-IID?
+    sampler: iid
+
+    # The random seed for sampling data
+    random_seed: 1
+
+trainer:
+    # The type of the trainer
+    type: basic
+
+    # The maximum number of training rounds
+    rounds: 5
+
+    # Whether the training should use multiple GPUs if available
+    parallelized: false
+
+    # The maximum number of clients running concurrently
+    max_concurrency: 1
+
+    # The target accuracy
+    target_accuracy: 0.94
+
+    # Number of epoches for local training in each communication round
+    epochs: 5
+    batch_size: 32
+    optimizer: SGD
+    learning_rate: 0.01
+    momentum: 0.9
+    weight_decay: 0.0
+
+    # The machine learning model
+    model_name: lenet5
+
+algorithm:
+    # Aggregation algorithm
+    type: fedavg

examples/federated_learning/surface_defect_detection/aggregation_worker/server.yml

@@ -6,7 +6,7 @@ clients:
     total_clients: 0
 
     # The number of clients selected in each round
-    per_round: 1
+    per_round: 2
 
     # Should the clients compute test accuracy locally?
     do_test: true

examples/federated_learning/surface_defect_detection/training_worker/config.yml

@@ -0,0 +1,63 @@
+clients:
+    # Type
+    type: simple
+
+    # The total number of clients
+    total_clients: 3
+
+    # The number of clients selected in each round
+    per_round: 1
+
+    # Should the clients compute test accuracy locally?
+    do_test: true
+
+server:
+    address: 127.0.0.1
+    port: 7363
+
+data:
+    # The training and testing dataset
+    datasource: MNIST
+
+    # Where the dataset is located
+    data_path: ./data
+
+    # Number of samples in each partition
+    partition_size: 20000
+
+    # IID or non-IID?
+    sampler: iid
+
+    # The random seed for sampling data
+    random_seed: 1
+
+trainer:
+    # The type of the trainer
+    type: basic
+
+    # The maximum number of training rounds
+    rounds: 5
+
+    # Whether the training should use multiple GPUs if available
+    parallelized: false
+
+    # The maximum number of clients running concurrently
+    max_concurrency: 1
+
+    # The target accuracy
+    target_accuracy: 0.94
+
+    # Number of epoches for local training in each communication round
+    epochs: 5
+    batch_size: 32
+    optimizer: SGD
+    learning_rate: 0.01
+    momentum: 0.9
+    weight_decay: 0.0
+
+    # The machine learning model
+    model_name: lenet5
+
+algorithm:
+    # Aggregation algorithm
+    type: fedavg