Skip to content

A library for encrypted, privacy preserving machine learning

License

Notifications You must be signed in to change notification settings

mars-college/PySyft

 
 

Repository files navigation

Introduction

Binder Build Status Chat on Slack FOSSA Status

PySyft is a Python library for secure and private Deep Learning. PySyft decouples private data from model training, using Federated Learning, Differential Privacy, and Multi-Party Computation (MPC) within the main Deep Learning frameworks like PyTorch and TensorFlow. Join the movement on Slack.

PySyft in Detail

A more detailed explanation of PySyft can be found in the white paper on arxiv

PySyft has also been explained in videos on YouTube:

Pre-Installation

Optionally, we recommend that you install PySyft within the Conda virtual environment, for its simplicity in installation. If you are using Windows, we suggest installing Anaconda and using the Anaconda Prompt to work from the command line.

conda create -n pysyft python=3
conda activate pysyft # some older version of conda require "source activate pysyft" instead.
conda install jupyter notebook

Another alternative is to use python venvs. Those are our preferred environments for development purposes. We provide a direct install instructions in our makefile.

make venv

Installation

PySyft supports Python >= 3.6 and PyTorch 1.3

pip install syft[udacity]

This will auto-install the PyTorch and TF Encrypted dependencies, which are required for running the tutorials from Udacity's "Secure & Private AI" course (recommended).

You can install syft without these dependencies with the usual pip install syft, but you will need to install framework dependencies (i.e. PyTorch, TensorFlow, or TF Encrypted) yourself. If you feel you've received an unexpected installation error related to PyTorch or TF Encrypted, please open an issue on Github or reach out to #team_pysyft in Slack.

If you have an installation error regarding zstd, run this command and then re-try installing syft.

pip install --upgrade --force-reinstall zstd

If this still doesn't work, and you happen to be on OSX, make sure you have OSX command line tools installed and try again.

If this still fails, and you are on a Conda environment. It could be because conda provides its own compiler and linker tools which might conflict with your system's. In that case we recommend to use a python venv and try again.

You can also install PySyft from source on a variety of operating systems by following this installation guide.

Run Local Notebook Server

All the examples can be played with by running the command

make notebook

This assumes you want to use a local virtual environment. It installs it independently to the conda environment in case you installed one, or any other virtual environment you might have set up.

Once the jupyter notebook launches on your browser select the pysyft kernel.

Use the Docker image

Instead of installing all the dependencies on your computer, you can run a notebook server (which comes with Pysyft installed) using Docker. All you will have to do is start the container like this:

$ docker container run openmined/pysyft-notebook

You can use the provided link to access the jupyter notebook (the link is only accessible from your local machine).

NOTE: If you are using Docker Desktop for Mac, the port needs to be forwarded to localhost. In that case run docker with: bash $ docker container run -p 8888:8888 openmined/pysyft-notebook to forward port 8888 from the container's interface to port 8888 on localhost and then access the notebook via http://127.0.0.1:8888/?token=...

You can also set the directory from which the server will serve notebooks (default is /workspace).

$ docker container run -e WORKSPACE_DIR=/root openmined/pysyft-notebook

You could also build the image on your own and run it locally:

$ cd docker-images/pysyft-notebook/
$ docker image build -t pysyft-notebook .
$ docker container run pysyft-notebook

More information about how to use this image can be found on docker hub

Try out the Tutorials

A comprehensive list of tutorials can be found here

These tutorials cover how to perform techniques such as federated learning and differential privacy using PySyft.

High-level Architecture

alt text

Start Contributing

The guide for contributors can be found here. It covers all that you need to know to start contributing code to PySyft in an easy way.

Also join the rapidly growing community of 5000+ on Slack. The slack community is very friendly and great about quickly answering questions about the use and development of PySyft!

Troubleshooting

We have written an installation example in this colab notebook, you can use it as is to start working with PySyft on the colab cloud, or use this setup to fix your installation locally.

Organizational Contributions

We are very grateful for contributions to PySyft from the following organizations!

Udacity coMind Arkhn Dropout Labs

Disclaimer

Do NOT use this code to protect data (private or otherwise) - at present it is very insecure. Come back in a couple months.

License

Apache License 2.0

FOSSA Status

About

A library for encrypted, privacy preserving machine learning

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 99.8%
  • Other 0.2%