Secretflow - A Unified Framework For Privacy-Preserving Data Analysis And Machine Learning

SecretFlow is a unified framework for privacy-preserving data intelligence and machine learning. To achieve this goal, it provides:

  • An abstract device layer consists of plain devices and secret devices which encapsulate various cryptographic protocols.
  • A device flow layer modeling higher algorithms as device object flow and DAG.
  • An algorithm layer to do data analysis and machine learning with horizontal or vertical partitioned data.
  • A workflow layer that seamlessly integrates data processing, model training, and hyperparameter tuning.


For users who want to try SecretFlow, you can install the current release from pypi. Note that it requires python version == 3.8, you can create a virtual environment with conda if not satisfied.

pip install -U secretflow

Try you first SecretFlow program

>>> import secretflow as sf>>> sf.init(['alice', 'bob', 'carol'], num_cpus=8, log_to_driver=True)>>> dev = sf.PYU('alice')>>> import numpy as np>>> data = dev(np.random.rand)(3, 4)>>> data<secretflow.device.device.pyu.PYUObject object at 0x7fdec24a15b0>

Getting started


Contribution guide

For developers who want to contribute to SecretFlow, you can set up an environment with the following instruction.

git clone https://github.com/secretflow/secretflow.git# optionalgit lfs installconda create -n secretflow python=3.8conda activate secretflowpip install -r dev-requirements.txt -r requirements.txt

Coding Style

We prefer black as our code formatter. For various editor users, please refer to editor integration. Pass -S, --skip-string-normalization to black to avoid string quotes or prefixes normalization.


Non-release versions of SecretFlow are prohibited to use in any production environment due to possible bugs, glitches, lack of functionality, security issues or other problems.

Disqus Comments