Use the codes in this repository to simulate Orbital Learning platform.
Steps:
- Dowload the PTB-XL ECG dataset. Divide and create subsets of data representing various client datasets.
- Use OL_experiments/OL_experiments.py to adjust/connect data generated by clients and simulate scenarios such as varied melioration rate or tampering.
- Create a docker container using OL_experiments. Run it with volume connected to the folder containing your client datasets.
- Run OL_experiments/main.py
- Run OL_main.py Alter Orbital_system.py and Orbital_plane.py to alter thresholds or number of meliorations.
*** For LLM models, Please Use 'orbitallearning/client_llm' and 'orbitallearning/central_node_llm' docker images. These were used for the scalability demo.