installation.md

Prerequisites

Before you begin, ensure that you have the following prerequisites installed on your system:

• Python 3.6 or higher
• pip

Installing HeFlwr

Use pip to install the HeFlwr package by running the following command:

pip install heflwr

Installing Flower and PyTorch

For the full functionality of HeFlwr, it is essential to have Flower, PyTorch and Psutil installed. Follow these steps to install:

1. Install Flower using pip:

   pip install flwr

   If you also want to enable Flower's simulation features, you should run the following command:

   pip install flwr[simulation]

2. Current research on heterogeneous federated learning mainly focuses on traditional computer vision tasks, therefore we recommend installing PyTorch and torchvision:

   pip install torch torchvision

3. To better utilize the HeFlwr resource monitor, you also need to download Psutil:

   pip install psutil
   

Verifying the Installation

To verify that HeFlwr, Flower, PyTorch and Psutil have been successfully installed, you can run the following command in your terminal:

python -c "import heflwr; print(heflwr.__version__)"
python -c "import flwr; print(flwr.__version__)"
python -c "import torch; print(torch.__version__)"

These commands should return the version numbers of HeFlwr, Flower, PyTorch and Psutil, indicating that they have been successfully installed on your system.

Next Steps

After successfully installing HeFlwr, Flower, PyTorch and Psutil, you may want to:

• Visit HeFlwr's homepage for an overview of the design background and guiding principles.