Deep Learning Setup

In this page and the following subpages, I will record my experience on setting up my laptop for running deep learning.

1) Hardware requirement

First, I prefer a laptop for mobility. For true mobility, I prefer a light and thin laptop. To play with deep learning and train the network efficiently, I prefer a laptop with GPU. After some survey, I prefer MSI GS63VR laptop. I got it on sale during Black Friday, with the price of $1199, come with NVIDIA GeForce GTX 1060 6GB, 128 GB SSD, and 1TB HDD. After I got the computer, I have a clean install of the Windows 10, to eliminate the bloatware come with the laptop. After play with the computer for few days, I decided to upgrade the hard drive storage, as 128 GB SSD is too small for dual boot. I bought 500 GB Samsing 850 Evo PCIe SSD ($140), and 1TB Samsung 850 Evo SATA SSD ($280). Before changing the hard dive, I purchase two casing, one for the PCIe SSD (20x80mm), another one for 2.5" SATA disk. Samsung SSD comes with a disk cloning software, so I clone my Windows OS in 128 GB SSD to my new 500 GB SSD before I switch the drive. The steps in changing the hard drive are given in:

https://www.ultrabookreview.com/12757-upgrade-ram-ssd-msi-gs73/

it involves a lot of patient and careful, it took my the whole evening.

2) Linux OS installation - dual boot.

3) NVIDIA driver installation

The NVIDIA driver installation is easy, go to Additional Drivers,

select the NVIDIA binary driver and install.

The NVIDIA driver installation comes with NVIDIA PRIME, that enables us to choose either the Intel (save power) or the NVIDIA graphic card (for better performance).