Explore Available Examples

Training Keras, TensorFlow 2.1 and PyTorch models with different fusion algorithms:

• Running federated averaging (FedAvg)
• Simple average
• FedPlus with Keras
• Gradient aggregation
• PFNM with Keras
• Coordinate median with Keras
• Krum with Keras
• Zeno with Keras

Training scikit-learn models in IBM FL:

• Logistic classifier
• SPAHM with KMeans
• Differential private Naive Bayes models

ID3 Decision trees:

• Training ID3 decision trees

Reinforcement learning in IBM FL:

• Cartpole
• Pendulum

Unsupervised learning in IBM FL:

• SPAHM with KMeans

Jupyter Notebooks to run IBM FL:

• Keras Classifier
• Reinforcement learning Cartpole

IBM FL Command Reference

IBM FL Command	Participant	Description
START	aggregator / party	Start accepting connections
REGISTER	party	Join an FL project
TRAIN	aggregator	Initiate training process
SYNC	aggregator	Synchronize model among parties
STOP	aggregator	End experiment process
EVAL	party	Evaluate model