Introduction

This is a VirtualBox VM that is automatically generated using Vagrant.

Machine Learning and Deep Learning packages installed are: Scikit-Learn, NLTK, Keras, TensorFlow and Theano. A Vagrant file is used to generate this VM, which runs on Ubuntu 14.04 (trusty64).

Getting Started

• This VM should work in Windows, macOS and Linux

• VirtualBox (version 6+) is required

• Download and install Vagrant

• Clone the virtual machine specs with:

git clone https://github.com/f0nzie/vagrant-deeplearning-t64shjpythtfkeskl-apy352.git

• Change to the folder where this repo has been cloned and type on your local terminal:

  vagrant up
  

• Browse to Jupyter with: http://127.0.0.1:8100/. Try the different notebooks in there.

• To access the virtual machine console or terminal, type:

  vagrant ssh
  

• When finished, power off the virtual machine with:

  vagrant halt
  

What’s Installed

• Deep Learning
  • Keras 2.0.6
  • TensorFlow 1.3.0
  • Theano 0.8.2
• Machine Learning
  • sklearn 0.18
  • nltk 3.2.1
• Python
  • Anaconda3 Python 3.5.2
  • numpy 1.11.1
  • scipy 0.18.1
  • matplotlib 1.5.3
  • bokeh 0.12.2
  • pandas, 0.18.1
  • seaborn 0.9.0
  • jupyter 1.0.0
  • jupyter_core 4.2.0
  • ipython 5.1.0
  • h5py 2.6.0
  • pip 8.1.2

Jupyter notebook server

Jupyter notebook server is available at the host’s browser at http://localhost:8100.

Password : password

Testing

There are several notebooks to test the deep learning packages:

• test_keras.ipynb
• test_tensorflow.ipynb
• theano_test.ipynb
• test_loop_packages.ipynb: check the versions of the main installed packages

Other

• Disk size: 40 GB
• RAM: 4096 GB
• CPUs: 2
• Network: NAT
• USB: off
• Shared folder: 2

Codes machine name

• t64: Ubuntu Trusty 64-bit
• sh: shell script provisioning
• jpy: Jupyter server
• rss: RStudio Server
• th: Theano
• tf: TensorFlow
• ke: Keras
• skl: scikit-learn
• -apy: Anaconda Python followed by version number
• Host name: T64G40A8100
• VM name: vagrant-T64G40DLSJ352A8100
• Short name: T64G40DLSJ352A8100

Notes

• There is only one version of Anaconda3 that works with the setting in the provisioning files, and that is Anaconda3-4.2.0-Linux-x86_64.sh. Changing it to 4.3.0 or 4.4.0 will not allow to auto start Jupyter.