Machine Learning Environment Setup

Machine learning is the hottest topic in the field of technology right now. We can apply machine learning everywhere in our daily lives and surroundings. We will learn in this article how to setup the environment for machine learning, so let’s begin.

PC Hardware Setup

To perform machine learning or deep learning on any given dataset with vast amounts of valuable data, we need to ensure the PC is strong enough to handle the computations.

1. CPU- Central Processing Unit – We will require an Intel i5 6th generation process or higher.

2. RAM – The minimum requirement is 8GB and can go up to 16GB.

3. GPU- Graphics Processing Unit – NVIDIA GEFORCE GTX90 or higher. AMD GPUs are unable to perform deep learning.

4. OS- Operating System – It is preferable to choose Ubuntu or Windows 10 or higher.

Downloading and Setting up

Step 1: Downloading Anaconda

First, we will download anaconda package with python for your software platform.
Anaconda is an open-source and free platform for developing ML algorithms using python.
You can click on the link given to download the 32 or 64-bit version and get started:

Download the software based on your version. Here we have taken Windows, but you can take one based on your OS.

Step 2: Installing Anaconda

After you download the setup, it is easy to install Anaconda. You just need to open the wizard and click on the next instructions on the screen.

It takes a few minutes for the initial set-up.

Step 3: Updating Anaconda

Search for the Anaconda prompt and type the commands given below:

conda update conda
conda update –all

Step 4: Downloading CUDA Toolkit and cuDNN

1. Click on the link to download CUDA Toolkit version 8.0 or 9.0.

2. Download cuDNN

Install the current version of cuDNN. Then, choose your version and CUDA based on your operating system. Finally, create an account with the help of your email id and become a member to access the software.

Step 5: Adding cuDNN in the Environment Path

It is a crucial step. So make sure you follow it carefully.

a. Open the run dialogue with the help of (Win + R) and run the following command: sysdm.cpl.

b. Go to system properties in windows 10, and click on the ‘Advanced’ tab.

c. Click on the environment variables.

d. Type the following path environment :

C: \cudnn-9.0-windows 10-x64-v7\cuda\bin

Step 6: Making an environment for Anaconda

Coming to the last stages in setting up Anaconda, let us try to create an environment in Anaconda.
We will make an Anaconda environment for every different usage. So it doesn’t interrupt the initial environment.

a. Make a conda environment with the name “tensor” (this could be any name):

Conda create -n tensor pip python =3.6

b. Type in the below command to activate the environment:

Activate tensor
(tensor) C:>

Step 7: Installing Deep Learning Libraries

Here we will download all the relevant deep learning libraries required to begin your deep learning journey.

a. Tensorflow

It is an open-source machine learning framework for powering deep neural networks with high-level code. The Google Brain team developed it and released it in 2015. It is most commonly used with python, but it can run in any other programming language.

Type the following code to install the tensor flow in the Anaconda prompt:

C:\ pip install tensor-gpu

or

C:\ pip install tensor-CPU

b. Keras

Keras is a high-level deep learning API written in python for neural networks. It supports multiple backend neural network computations and makes implementing neural networks fairly easy.
Open your anaconda prompt and type the below command.

pip install Keras

Conclusion

This article covered how to set up the environment for installing and setting up the environment for machine learning.