Skip to content

Latest commit

 

History

History
278 lines (184 loc) · 9.5 KB

Installation.md

File metadata and controls

278 lines (184 loc) · 9.5 KB

Vocabulary

  • open app means to open/execute the app, usually by pressing HOME/WINDOWS Key and searching for app
  • command lines (code) are meant to be written in a terminal. To open a terminal, either open terminal or press Ctrl + Alt + T

Install Ubuntu

Make a bootable USB Drive from a Ubuntu computer (or simply ask me for one)

  1. Download the latest stable Ubuntu Desktop
  2. Have an empty USB Stick
  3. Open Startup Disk Creator (if not on Ubuntu, find an equivalent)
  4. Fill in the fields as follows and press Make Startup Disk
    • Source disc image: the .iso file you just downloaded
    • Disk to use: your USB stick

Booting on that drive

If your disk is encrypted by windows, you may need to unencrypt it (deactivate BitLocker) to install Ubuntu (this depends on what you do). There is however no risk in just continuing from here without caring about this yet, except losing time.

Furthermore, installing Ubuntu has a very low chance of going wrong, but if it goes wrong, you may lose all your data. It may be wise to backup your data beforehand...

  1. Make sure to press F2 or F12 (or sometimes F8, this is depends on your hardware) during startup and select your USB Stick as boot device
  2. Select Try or Install Ubuntu
  3. Select Try Ubuntu (do not install yet)
  4. Try as much as you want

Creating space to install Ubuntu

  1. Open Disks from the ubuntu you are trying
  2. Make a partition for where you want to install Ubuntu. It may be impossible without erasing the whole disk if the disk in encrypted.

Installing Ubuntu

  1. Press the Install Ubuntu Icon on the top-right
  2. Go through the questions, selecting your choice for the question you know how to answer, making sure to select the correct partition and leaving the defaults for the other questions.

Setup system configuration

Open a terminal either by pressing Ctrl + Alt + T or by pressing Home/Windows key and searching for terminal.

Install git, htop and curl

Installing git allows you to fetch repositories from github and is thus the start of our installation process. htop is a simple command to show CPU/Memory usage in terminal. Curl is a simple command line download tool.

sudo apt install git
sudo apt install htop
sudo apt install curl

Creating basic folder structure for installations

We suggest you put all installation material in an CustomSetupMaterial folder in your home directory. To do so:

cd ~ #Goes to your home directory
mkdir CustomSetupMaterial #Creates Installs folder
cd CustomSetupMaterial #Goes inside it

Downloading this repo

Once you are in CustomSetupMaterial (last command of last step) simply execute:

git clone "https://github.com/JulienBrn/DefaultInstall.git"

Install zsh (Shell configuration)

Basic zsh configuration

Warning: You may not see the results before rebooting

sudo apt install zsh   # Installs zsh
chsh -s $(which zsh)   # Sets zsh as default shell

Adding oh-my-zsh configuration

Fetches oh-my-zsh install file and launches it (thus installing the git repository in ~/.oh-my-zsh and setting ~/.zshrc to use it (i.e. use source command on it).

cd ~
sh -c "$(wget https://raw.githubusercontent.com/ohmyzsh/ohmyzsh/master/tools/install.sh -O -)"

Configure side-bar

You can play with the preferences settings of the GUI or simply run

dconf load / < ~/CustomSetupMaterial/DefaultInstall/side_bar.dconf

Install Guake (Terminal configuration)

Basic guake install

Guake is a drop-down terminal that is simple to use with many features.

sudo add-apt-repository ppa:linuxuprising/guake #adds up to date repo
sudo apt-get update
sudo apt install guake

Configure guake

This step configures Guake with good default behavior. Mainly it binds F12 to toggling, defines Ctrl + | or Ctrl + - for splitting terminals, ... The easiest way is simply to run the following command and can be done through the graphical interface with settings -> keyboard -> view and customize shortcuts -> custom shortcuts (for F12 binding to command guake -t) and opening Guake Preferences and configure shortcuts.

dconf load / < ~/CustomSetupMaterial/DefaultInstall/guake_settings.dconf

Run Guake

Run Guake now

Run Guake now by pressing Alt + F2 and typing guake

Set Guake as start-up application

Simply run

cp ~/CustomSetupMaterial/DefaultInstall/guake_autostart.txt ~/.config/autostart/guake.desktop

OR

Open startup application manager. Press Add and fill the following fields:

  • name: Guake (or whatever you want)
  • Command: guake
  • Comment: Starts Guake terminal at startup (or whatever you want)

Setup low memory warning script

This is a simple script that warns you when the memory is low allowing you to act before the computer becomes slow.

Simply run

cp ~/CustomSetupMaterial/DefaultInstall/memory_autostart.txt ~/.config/autostart/memory.desktop
sed -i "s;~;$HOME;" ~/.config/autostart/memory.desktop

OR

Open startup application manager. Press Add and fill the following fields:

  • name: Memory Warning (or whatever you want)
  • Command: /home/user/CustomSetupMaterial/DefaultInstall/low_memory_warning.sh where you replace user by your current user name.
  • Comment: Warns on low memory (or whatever you want)

Accessing the Filer

I created a program for this specific purpose and normally you should only need to follow the instructions. To execute this program, run

sudo python3 ~/CustomSetupMaterial/DefaultInstall/install_filer.py

Setting up your Github key

Setup Email Configuration

Setup Firefox Configuration

The suggestion is simply to use a good existing profile. Documentation on how to do that to come.

Setup Programming Environment

Setup a Python programming environment

I suggest to use python within a conda environment. This allows you to separate your python installation for your programs from the system python installation. It also enables you to create several nearly fully isolated environments (unline other virtual envrinments tools).

Installing a conda environment

Install miniconda using:

cd ~/Downloads #Puts you in the download directory
curl https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -o Miniconda3-latest-Linux-x86_64.sh #Downloads the latest version of miniconda
bash Miniconda3-latest-Linux-x86_64.sh #Launches install
#Accept the LICENCE and say yes when prompted to update shell profile
source ~/.zshrc #Take into account the changes

Create a conda environment:

conda create --name CONDA_ENV_NAME python=3.11 #Change the name (CONDA_ENV_NAME) to what you want to call the environment. Set the python version to was is right for you
conda activate CONDA_ENV_NAME # Activate the environment

Play with the python interpreter:

python
#Run python commands such as 3+5
#Press CTRL+D to exit

Setting up the VSCode IDE for Python

Go to the linux vscode webpage and download the .deb file. Just click on the downloaded file and select software install or run the following

sudo apt install FILE.deb #where FILE.deb is the file you just downloaded. You may need to cd to the Downloads directory

Open the vscode IDE, then either sync it with one of your previous accounts (whose settings probably already include Python), or click on the extension icon on the left, search for python and install the Python language support extension.

Finally, got to file -> new file and choose python file.

Write the following program in the file and save it:

print(f"Hello! Did you know that 3+5={3+5}?")

In the lower right, click on Python and select the your interpreter (you should choose the one of your conda environment)

Finally, press the "play" button and see what happens!

Setting up the VSCode IDE for Jupyter

Simply create a Jupyter notebook file and follow the instructions given by vscode

Enabling GPU accelaration for tensorflow (guide inspired from tensorflow guide)

Check that the following command runs and provides a reasonable output. If not please enable GPU accelaration following some other guide.

nvidia-smi

Install required packages. Make sure your conda environment (not base) is activated.

conda install -c conda-forge cudatoolkit=11.8.0
pip install nvidia-cudnn-cu11==8.6.0.163

Configure environment variables

mkdir -p $CONDA_PREFIX/etc/conda/activate.d
echo 'CUDNN_PATH=$(dirname $(python -c "import nvidia.cudnn;print(nvidia.cudnn.__file__)"))' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
echo 'export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$CONDA_PREFIX/lib/:$CUDNN_PATH/lib' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh

Install tensorflow

pip install --upgrade pip
pip install tensorflow

Check tensorflow installation

python3 -c "import tensorflow as tf; print(tf.reduce_sum(tf.random.normal([1000, 1000])))"
# Should output tf.Tensor(Some_num, shape=(), dtype=floatxx)
# Possibly with lots of prints before

Check that GPU is enabled

python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"
# Should return a list with at least one element such as: [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]