Update simulation step-by-step example

examples/flower-simulation-step-by-step-pytorch/Part-I/client.py

@@ -98,7 +98,7 @@ def client_fn(cid: str):
             trainloader=trainloaders[int(cid)],
             vallodaer=valloaders[int(cid)],
             num_classes=num_classes,
-        )
+        ).to_client()
 
     # return the function to spawn client
     return client_fn

examples/flower-simulation-step-by-step-pytorch/Part-II/client.py

@@ -75,6 +75,6 @@ def client_fn(cid: str):
             trainloader=trainloaders[int(cid)],
             vallodaer=valloaders[int(cid)],
             model_cfg=model_cfg,
-        )
+        ).to_client()
 
     return client_fn

examples/flower-simulation-step-by-step-pytorch/README.md

@@ -1,5 +1,7 @@
 # Flower Simulation Step-by-Step
 
+> Since this tutorial (and its video series) was put together, Flower has been updated a few times. As a result, some of the steps to construct the environment (see below) have been updated. Some parts of the code have also been updated. Overall, the content of this tutorial and how things work remains the same as in the video tutorials.
+
 This directory contains the code developed in the `Flower Simulation` tutorial series on Youtube. You can find all the videos [here](https://www.youtube.com/playlist?list=PLNG4feLHqCWlnj8a_E1A_n5zr2-8pafTB) or clicking on the video preview below.
 
 - In `Part-I` (7 videos) we developed from scratch a complete Federated Learning pipeline for simulation using PyTorch.
@@ -19,20 +21,17 @@ As presented in the video, we first need to create a Python environment. You are
 # I'm assuming you are running this on an Ubuntu 22.04 machine (GPU is not required)
 
 # create the environment
-conda create -n flower_tutorial python=3.8 -y
+conda create -n flower_tutorial python=3.9 -y
 
 # activate your environment (depending on how you installed conda you might need to use `conda activate ...` instead)
 source activate flower_tutorial
 
 # install PyToch (other versions would likely work)
 conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.6 -c pytorch -c nvidia -y
-# conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 -c pytorch # If you don't have a GPU
+# conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 -c pytorch -y # If you don't have a GPU
 
-# install flower (for FL) and hydra (for configs)
-pip install flwr==1.4.0 hydra-core==1.3.2
-# install ray
-# you might see some warning messages after installing it (you can ignore them)
-pip install ray==1.11.1
+# Install Flower and other dependencies
+pip install -r requirements.txt
 ```
 
 If you are running this on macOS with Apple Silicon (i.e. M1, M2), you'll need a different `grpcio` package if you see an error when running the code. To fix this do:

examples/flower-simulation-step-by-step-pytorch/requirements.txt

@@ -0,0 +1,2 @@
+flwr[simulation]>=1.0, <2.0
+hydra-core==1.3.2