Here we provide tutorials on running FATE jobs:
Submitting jobs with Pipeline is recommended, here are some Jupyter Notebook with guided instructions:
- Upload Data with FATE-Pipeline
- Train & Predict Hetero SecureBoost with FATE-Pipeline
- Build NN models with FATE-Pipeline
- Upload & Train Hetero SecureBoost on Data with Match ID
- Upload & Train Hetero SecureBoost on Data with Meta
- Upload & Run An Intersection Task on Data with Multiple Match IDs
Submitting jobs without Pipeline is supported as well, which one should provide job configuration(s) in json format:
Models can be published with FATE Serving to Serving without FATE:
And for those who want to run jobs in batches, ie. run algorithm tests, try using fate_test:
To merge models from different roles and export as sklearn/LightGBM or PMML format, please refer to the tutorial below: