Flower version: 1.9.0
Python 3: 3.8.0
Torch: 2.3.1
Torchvision: 0.18.1
Ubuntu: 18.04
virtualenv : 20.17.1
pip: 9.0.1(old) 24.1(new)
Setup tools: 39.0.1 (old) 68.0.0(new)
tensorflow: 2.16.1 (pi and iMac) (2.13.1 Jetson, to be determine if it should be for all device)
numpy: 1.26.4 (pi and iMac) 1.24.4 Jetson
pandas: 2.2.2 (pi and iMac) (unsure for jetson, checking process stoped due to issue with installing tensorflow)
Protobuf - 3.20.3
Iterators - 0.0.2
ptotobuf - 4.25.2 (jetson)
matplotlib: 3.7.5 (jetson)
grpcio: 1.64.1
python3 -V (confirm below 3.8.0)
sudo apt update
sudo apt install python3.8
sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.6 1
sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.8 2
sudo update-alternatives --config python3
2
python3 -V (check if python3 is now version 3.8.0)
python3 -m pip install --upgrade pip (make sure pip is 24.1)
pip install flwr
To run img_clsf,
pip install torch torchvision
Then cd to FL_Project and start training
To run ids_dnp3
for Jetson
pip install pandas
If you have encounter issues with "ImportError: C extension: No module named 'pandas._lib.tslib 'not built...." when running python3.server.py
Try
sudo pip3 uninstall pandas
sudo pip3 install pandas
If scikit is a issue, try
pip install --upgrade scikit-learn
pip install tensorflow
If HDF5 is a issue, try
To download HDF5 in version 1.12
And make h5py version 3.10.0
_ If Wheel is a issue, try
In window,
To find your IP address, follow the steps.
On the Windows machine, click Start. In Search, enter “command prompt” and then click Enter.
In the command prompt, enter “ipconfig/all” and click Enter.
Your IP address appears as an IPv4 address under Ethernet adapter Ethernet.
In Linux/Ubuntu/Pi OS,
To find your IP address on the command line, run:
“ifconfig”
To get a list of IP addresses, use the ip command. For example:
“ip -4 -o a | cut -d ' ' -f 2,7 | cut -d '/' -f 1”
Use port 8080
If “No module named ‘pip._internal’
Python3 -m pip install –upgrade pip
If “ModuleNotFoundError: No module named 'flwr'”
pip install flwr-datasets
python3 server.py
To check flwr version, use “pip show flwr”
To exit the virtual environment, use “deactivate”
To exit the dir, use “cd”
To access the venv in FL_Project
cd FL_Project
source Project_env/bin/activate
pip install torch torchvision
If Jetbot doesn't have enough space,
sudo apt-get install gparted
manually config the SD card size to it's max
https://tech.serhatteker.com/post/2019-09/upgrade-python37-on-ubuntu18/
https://www.youtube.com/watch?v=DhLu8sI9uY4
https://flower.ai/docs/examples/embedded-devices.html#setting-up-a-jetson-xavier-nx