Table 3.
Comparison of the open-source federated learning systems’ (FLSs’) features.
| Features | TFF 0.17.0 | FATE 1.4.4/1.5 | PySyft 0.2.8 | PFL 1.1.0 | FL & DP 0.1.0 |
|---|---|---|---|---|---|
| OS | Mac Linux |
Mac Linux |
Mac Linux Win iOS Android |
Mac Linux Win |
Linux Win |
| Settings | Cross-silo | Cross-silo | Cross-silo Cross-devices |
Cross-silo Cross-devices (in future) |
Cross-silo |
| Data Partitioning |
Horizontal | Horizontal Vertical |
Horizontal Vertical |
Horizontal Vertical Transfer |
Horizontal |
| Data type | Time series Images |
Time series | Images | Time series Images |
Time series Images |
| Mode | Simulation | Simulation Federated |
Simulation Federated |
Simulation Federated |
Simulation |
| Clustering model | No | No | No | No | Yes (Kmeans SciKitLearn) |
| ML Model | No | Yes (very slow) |
No | Yes | Yes (SciKitLearn) |
| Decision Tree Model |
No | Yes (very slow) |
No | No | No |
| Protocol | gRPC/proto (in future) |
gRPC/proto | Doesn’t use | ZeroMQ | Doesn’t use |