Skip to content

Decentralized learning using a Tangle architecture

Notifications You must be signed in to change notification settings

osmhpi/learning-tangle

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

85 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Federated Learning Tangle Experiment

Use pipenv to set up your environment.

Write the EMNIST dataset to disk using write_emnist_dataset.py.

Customize NUM_ROUNDS and NUM_CLIENTS in experiment.py and run it.

For executing a single step (i.e. a particular client training on local data and submitting a transaction), run step.py f0000_14 10, where f0000_14 is the client id and 10 corresponds to tangle_data/tangle_10.json.

To view tangles in a web browser, run python -m http.server in the repository root and open http://localhost:8000/viewer/. You may need to slide the slider to the left to see something.