Setup instructions

1. Python3

You will need Python 3.6 for running example code.

To check presence of Python 3 on your system:

$ python3 -V
# Python 3.6.3

MacOS

MacOS usually comes with Python 2.7 preinstalled, but this is not sufficient.

If you don't have Python 3 yet, use brew for installation:

$ brew install python3

Also install pip3 and virtualenv:

$ sudo easy_install pip
$ pip install --upgrade virtualenv

Ubuntu

Ubuntu 16.04 and later come with both Python 2 and 3 preinstalled.

Additionally install pip package manager and virtualenv:

$ sudo apt-get install python3-pip python3-dev python-virtualenv

2. Clone deep-client repo

The next step is to clone deep-client repository to your computer:

$ git clone https://github.com/dimakura/deep-client.git
$ cd deep-client/

3. Environment setup

To create a new environment:

$ virtualenv --system-site-packages -p python3 .env

All the remaining steps should be done inside this new environment. To activate environment, issue the following command:

$ source .env/bin/activate

This should change command prompt to:

(.env) $

Ensure pip is installed:

(.env) $ easy_install -U pip

4. Install PyTorch

Please refer to PyTorch installation page in case of trouble.

MacOS

(.env) $ pip3 install http://download.pytorch.org/whl/torch-0.2.0.post3-cp36-cp36m-macosx_10_7_x86_64.whl

Ubuntu

(.env) $ pip3 install http://download.pytorch.org/whl/cu75/torch-0.2.0.post3-cp36-cp36m-manylinux1_x86_64.whl

5. Additional libraries

Install all other libraries using pip:

(.env) $ pip3 install torchvision
(.env) $ pip3 install --upgrade jupyter

Working with environment

To activate working environment:

$ source .env/bin/activate
(.env) $

You can quickly test PyTorch installation with:

(.env) $ python3 -c "import torch; print(torch.Tensor([1]))"
#
#  1
# [torch.FloatTensor of size 1]

When you are done working with it, issue:

(.env) $ deactivate
$

Working with Jupyter notebooks

To start jupyter notebooks:

(.env) $ jupyter notebook

and navigate to notebooks/ folder.