Machine Learning is not anymore a technology of the future, it’s a present-day miracle. We are using benefits of machine learning every now and then, maybe knowingly or maybe unknowingly. It has become a part of our daily life.

Today we are going to demonstrate a basic setup which will be required to run a machine learning application on your system.

Hardware Specification:

  • Processor: AMD Ryzen 1700X
  • Motherboard: Gigabyte
  • RAM: Corsair Vengeance LPX 16GB DDR4 2400MHz
  • GPU: nVidia GeForce GTX 1070 Ti
  • Hard Disk: 128GB SSD (PCIe)
  • SMPS: Coolermaster 650W
  • Operating System: Ubuntu 18.04 LTS 64-bit

In order to have a running setup for the machine learning & data science application, please follow the steps mentioned below.

  • Install latest Graphics driver 
    $ sudo add-apt-repository ppa:graphics-drivers
$ sudo apt update
$ sudo apt upgrade -y
$ sudo apt install nvidia-396 -y
  • Install CUDA 9.0 libraries 
    $ sudo apt install build-essential
$ wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_9.0.176-1_amd64.deb
$ sudo dpkg -i cuda-repo-ubuntu1604_9.0.176-1_amd64.deb
$ sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/7fa2af80.pub
$ sudo apt update
$ sudo apt install cuda-9-0
  • Install nVidia cuDNN for CUDA 9.0
    cuDNN is a proprietary package from NVIDIA, which provides support for GPU accelerated libraries for deep neural networks. In order to install cuDNN, you have to signup at https://developer.nvidia.com/cudnn and download the required package for CUDA 9.0 from here. Once the package is downloaded use following command to install it:

    $ sudo dkpg -i <<package-name>>.deb
  • Create basic development environment 
    $ sudo apt install python-pip python-virtualenv
$ virtualenv venv
$ source venv/bin/active
$ pip install tensorflow-gpu keras

This should give us a decent setup to start with a GPU accelerated machine learning application built on top of Tensorflow and Keras. Please feel free to get back to us with any queries or issues.