Model

• Synthetic data generation
• Train ML model
• Export ONNX network to input to EZKL

Workflow

Generate synthetic data

$ cargo run

Train Model and Save to ONNX Format

$ python src/model.py synthetic_data/credit_data.json

This generates the credit_model.onnx file.

Development Setup

Make sure that all the tools defined in .tool-versions are installed before you start.

Then install python dependencies:

$ pip install -r requirements.txt

Then build the project.

$ cargo build

Generate New Synthetic Data

If you want to re-train the model in new synthetic data:

$ cargo run

Then train the model and generate the new ONNX format:

$ python src/model.py synthetic_data/credit_data.json

Commit synthetic_data/credit_data.json and credit_model.onnx to git to make sure that we know which data have been used to train the particular model.

Note: The model is not evaluated for its correctness. This is something that we could have done with test data, but we haven't for now.