Skip to content

Latest commit

 

History

History
204 lines (108 loc) · 6.35 KB

INSTALLATION.md

File metadata and controls

204 lines (108 loc) · 6.35 KB
Raw

Installation Guide

Before you start following the tutorials, you will need to install PySyft on your computer.

Here are the PySyft installation instructions for Linux, Mac OS, and Windows.

Authors:

Linux

Installing PySyft on Linux is really straight forward. Here are the steps:

1. Make sure you have Python >= 3.6 < 3.8

2. Install PyTorch 1.4

Get the installation command here (use the pip option) and run it in the terminal.

NOTE: Use exact version 1.4.0, not just 1.4. E.g. pip install torch==1.4.0.

3. Clone the PySyft repo from Github

git clone https://github.com/OpenMined/PySyft.git

4. Enter the cloned repo

cd PySyft

5. Install PySyft

python setup.py install

6. (Optional)Install TF Encrypted dependencies

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

pip install -r pip-dep/requirements_udacity.txt

7. Test your installation

python setup.py test

Mac OS

1. Install Python

In order to install python on MacOs, you first need to install Homebrew the famous package manager.

Start by opening a terminal and type the following:

xcode-select --install
 /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"

This will install Homebrew. Then, install python (version >= 3.6 < 3.8) by running:

brew install python3

2. Install PyTorch 1.4

You should first get the installation command here (use the pip option)

Then run that command in the open terminal.

NOTE: Use exact version 1.4.0, not just 1.4. E.g. pip install torch==1.4.0.

3. Clone the PySyft repo from Github

git clone https://github.com/OpenMined/PySyft.git

4. Install PySyft

First, you need to enter the PySyft folder you cloned earlier:

cd PySyft

Then, run this to install the PySyft package:

python setup.py install

(Optional)Then install the TF Encrypted dependencies, which are required for running the tutorials from Udacity's "Secure & Private AI" course (recommended). Run:

pip install -r pip-dep/requirements_udacity.txt

You can test your installation by running:

python setup.py test

Windows

Currently, the PyTorch 1.4.0 package is not available in Windows, preventing an installation of PySyft from being completed successfully. A workaround is to grab an Ubuntu Submachine from the Windows store, and then following the instructions for the Linux installation.

1. Install Python

PySyft requires python version Python >= 3.6 < 3.8

Here is the link to install python www.python.org/downloads

2. Add python and pip to the PATH environment variable

First, you need to find the folder path to which python.exe and pip3 were installed.

In my case it is:

C:\Users\alan\AppData\Local\Programs\Python\Python37

for the former, and:

C:\Users\alan\AppData\Local\Programs\Python\Python37\Scripts

for the latter. So I will add both paths to the PATH environment variable.

In order to do that, go to Run and type sysdm.cpl and press Enter. The following window should open:

Click on the Environment Variables... button in the bottom-right corner. The following window should appear:

Select the Path row in the System varibles section as in the above screenshot and click on Edit...

In the window that opens, click on New and add the installation paths for python.exe and pip. Here is an example:

Click on OK.

3. Install PyTorch 1.4

You should first get the installation command here (use the pip option)

NOTE: Use exact version 1.4.0, not just 1.4. E.g. pip install torch==1.4.0.

Then open the command prompt by going to Run and typing cmd, type or paste the pytorch installation command you got from the above link and press Enter.

4. Install Git

Here is the download link for Git on windows: git-scm.com/download/win

5. Install Microsoft Build tools

Go to the download page and click on Free download under Community in the Visual Studio download section. This is illustrated in the following screenshot:

After the download is finished, run the downloaded package. You will eventually get the following window:

Select Desktop development with C++ and click on Install at the bottom-right corner of the page. (In the above screenshot you see a Close button instead since I have already installed it.)

Wait until the the install has finished.

6. Clone the PySyft repo from Github

git clone https://github.com/OpenMined/PySyft.git

7. Install PySyft

First, you need to enter the PySyft folder you cloned earlier:

cd PySyft

Then

python setup.py install

This will install PySyft and its requirements to your system. You should see no errors.

(Optional)Then install the TF Encrypted dependencies, which are required for running the tutorials from Udacity's "Secure & Private AI" course (recommended).

pip install -r pip-dep/requirements_udacity.txt

You can also test your installation by running:

python setup.py test

Troubleshooting

If installation is not working for you, please tell us about your issue on our slack channel (slack.openmined.org)

We will do our best to help!!