实验环境配置

Pre-requisites: git, python 2.7.X, virtualenv, pip (7.1.X recommended)

• If you're using Ubuntu, here are all the packages you'll need before you can proceed

  $ sudo apt-get install python2.7 python-dev build-essential curl libatlas-base-dev gfortran
  $ sudo apt-get install libfreetype6-dev libpng-dev libjpeg-dev
  

• Clone the repo from GitHub

  $ git clone https://github.com/rouseguy/intro2deeplearning.git
  $ cd intro2deeplearning
  

• Create python virtual environment

  $ virtualenv env
  $ source env/bin/activate
  

• Install requirements using pip

  $ pip install -r requirements.txt
  

  Use requirements_linux.txt instead of requirements.txt if you're on linux

• When the requirements are being downloaded / installed, Fetch the datasets simultaneously

    $ sh download_data.sh
  

• Run check_env.py script to test the dependencies

  $ python check_env.py
  

  Output should look like this

  [ OK ] scipy version 0.15.1
  [ OK ] PIL version 1.1.7
  [ OK ] keras
  [ OK ] IPython version 4.0.0
  [ OK ] theano version 0.7.0
  [ OK ] numpy version 1.9.2
  [ OK ] pandas version 0.16.2
  [ OK ] gensim version 0.10.3
  [ OK ] sklearn version 0.16.1
  

  This means you have all the dependencies installed and you're ready to start.

• Run the notebook

  $ cd notebooks
  $ ipython notebook
  

  This opens your default browser which displays the list of notebooks in the current directory. 

•