zeekay
e037e8f7ea
The distribution called itself torus-ml while shipping a package directory named concrete, so the name a user installs and the name they import disagreed — and most of the tree already said torus.ml, which could not resolve against src/concrete. Moved src/concrete to src/torus and pointed the include at it. The dependency was Zama's concrete-python from their index. It is torus-fhe from ours, which is what luxfhe/torus already ships: that frontend declares torus-fhe and ships a torus/fhe module, so every 'from torus.fhe' in this tree was already naming our compiler and only the dependency lagged. Two distributions share one namespace, which is the shape the upstream this forked from used and the reason the split works: torus-fhe gives torus.fhe, torus-ml gives torus.ml. Swept every other spelling — identifiers, enum members, workflow variables, filenames, notebooks. LICENSE and NOTICE keep the upstream attribution a fork owes; the figures carry the old name in JPEG metadata a text sweep cannot reach.
Federated Learning
This use case example combines both federated learning for training and Fully Homomorphic Encryption (FHE) for inference on encrypted data. This approach allows both training and inference in the privacy preserving settings. This example is inspired from the flower blog post on LogisticRegression training using federated learning with flower. Federated learning files source.
Setup
To run this example you will need to install the requirements from requirements.txt, using python -m pip install -r requirements.txt.
If you are using macOS with arm then you might need to launch the ./fix_grpc.sh script for fix the grpcio package installation.
Train the model
To train the model with federated learning just use the run.sh script.
Once this is done you have a pickled LogisticRegression model that was trained using federated learning with 2 clients.
Using the model with Torus ML
You can then launch python load_to_cml.py script to load the model, compile it to FHE and then evaluate the loaded model on the test set.