References:
1. OpenMined Documentation - PySyft example on Breast Cancer study: https://docs.openmined.org/en/latest/getting-started/introduction.html
Dependency/Module installations:
Install Jupyter Notebook and create a folder
Install Pysyft, pandas, numpy and sklearn libraries/modules using pip as below:
pip install syft
pip install pandas
pip install numpy
pip install sklearn
When running the Jupyter Notebook:
Restart kernel before running
Run one cell at a time and not all together as there are some initialization/configuration cells that have to be run only once, skip the next times you run the code again
Cell 5: Creating an admin account for each local site using default credentials Run this cell just once initially
Cell 7: Change the account credentials to make the site more secure Run this cell too just once, skip the next time you run the code again
**Cells 17 and 19: **Change the path that accesses the user dataset csv files as per your system before running
**Cell 99 **- Creating projects Run this cell once for initializing the projects and only when you run the code from scratch. Other times, access the project using the below code snippets: fitcheck_project1 = global_client1.projects[0] fitcheck_project2 = global_client2.projects[0]
**Cell 111: **Creating code requests Run this cell only once, unless you have restarted a Jupyter notebook and run the code from scratch. Otherwise skip. Running this multiple times creates multiple code requests which hinders sharing the results with the global model due to approval status:Pending.
Everything else has to be run._ _ If any errors occur especially when running cells with client initialization or database uploads, shut down and restart Jupyter Notebook and run the code from scratch.