running on multiple GPUs

Author: TsingZ0

README.md

@@ -182,11 +182,11 @@ For more details on datasets and FL algorithms in **IoT**, please refer to [FL-I
 cd ./dataset
 # Please modify train_ratio and alpha in dataset\utils\dataset_utils.py
 
-# python generate_MNIST.py iid - - # for iid and unbalanced scenario
-# python generate_MNIST.py iid balance - # for iid and balanced scenario
-# python generate_MNIST.py noniid - pat # for pathological noniid and unbalanced scenario
+python generate_MNIST.py iid - - # for iid and unbalanced scenario
+python generate_MNIST.py iid balance - # for iid and balanced scenario
+python generate_MNIST.py noniid - pat # for pathological noniid and unbalanced scenario
 python generate_MNIST.py noniid - dir # for practical noniid and unbalanced scenario
-# python generate_MNIST.py noniid - exdir # for Extended Dirichlet strategy 
+python generate_MNIST.py noniid - exdir # for Extended Dirichlet strategy 
 ```
 
 The command line output of running `python generate_MNIST.py noniid - dir`
@@ -324,6 +324,7 @@ conda env create -f env_cuda_latest.yaml  # Downgrade torch via pip if needed to
     ```bash
     cd ./system
     python main.py -data MNIST -m CNN -algo FedAvg -gr 2000 -did 0 # using the MNIST dataset, the FedAvg algorithm, and the 4-layer CNN model
+    python main.py -data MNIST -m CNN -algo FedAvg -gr 2000 -did 0,1,2,3 # running on multiple GPUs
     ```
 
 **Note**: It is preferable to tune algorithm-specific hyper-parameters before using any algorithm on a new machine. 

docs/data.html

@@ -240,11 +240,11 @@ <h3 id="exp-mnist">Examples for <strong>MNIST</strong> in the <em><strong>label
 # In ./dataset
 # Please modify train_ratio and alpha in dataset\utils\dataset_utils.py
 
-# python generate_MNIST.py iid - - # for iid and unbalanced scenario
-# python generate_MNIST.py iid balance - # for iid and balanced scenario
-# python generate_MNIST.py noniid - pat # for pathological noniid and unbalanced scenario
+python generate_MNIST.py iid - - # for iid and unbalanced scenario
+python generate_MNIST.py iid balance - # for iid and balanced scenario
+python generate_MNIST.py noniid - pat # for pathological noniid and unbalanced scenario
 python generate_MNIST.py noniid - dir # for practical noniid and unbalanced scenario
-# python generate_MNIST.py noniid - exdir # for Extended Dirichlet strategy</pre>
+python generate_MNIST.py noniid - exdir # for Extended Dirichlet strategy</pre>
 
                 <p>The command line output of running <code>python generate_MNIST.py noniid - dir</code></p>
 

docs/quickstart.html

@@ -197,7 +197,9 @@ <h2>Usage</h2>
                     <li>Finally, run the <code>main.py</code> with choosen configurations, FL algorithm, and hyperparameters:</li>
                     <pre>
 # In ./system
-python main.py -data MNIST -m CNN -algo FedAvg -gr 2000 -did 0 # using the MNIST dataset, the FedAvg algorithm, and the 4-layer CNN model</pre>
+python main.py -data MNIST -m CNN -algo FedAvg -gr 2000 -did 0 # using the MNIST dataset, the FedAvg algorithm, and the 4-layer CNN model
+python main.py -data MNIST -m CNN -algo FedAvg -gr 2000 -did 0,1,2,3 # running on multiple GPUs
+</pre>
                 </ol>
                 <strong>Note</strong> It is preferable to tune algorithm-specific hyper-parameters before using any algorithm on a new machine. 
             </section>