Skip to content

Latest commit

 

History

History
60 lines (46 loc) · 2.32 KB

README.md

File metadata and controls

60 lines (46 loc) · 2.32 KB

Bruno-DaSilva's Dotfiles

At the moment, this does not actually contained dotfiles, but just my current .zshrc and iterm themes. Long term, this will become a proper dotfiles setup!

Install instructions

  1. brew tap homebrew/cask-fonts
  2. brew install --cask font-meslo-lg-nerd-font
  3. brew install zplug
  4. brew install zsh
  5. ./setup_zsh.sh
  6. cp .zshrc ~/.zshrc
  7. You may want to copy any configs from your ~/.bashrc and ~/.bash_profile files, or source them from ~/.zshrc as they won't be automatically called in the new shell.
  8. Import the gruvbox-dark iterm colors in to your iterm2 profile. Set it as the default profile.
  9. Modify your iterm profile to use Meslo LG S DZ Regular Nerd Font Complete (not Monospaced). You can use any nerd font here (comes with extra symbols), but I prefer Meslo!
  10. Close/open your terminal!

For Ubuntu/GUI (non-ssh'ing servers) installs

  • You'll have to install zsh without brew.
  • GUI ONLY: gnome-terminal is used instead of iterm. For the same theme setup, use the gnome-terminal dump with:
    • dconf load /org/gnome/terminal/legacy/profiles:/ < gnome-terminal-profiles.dconf.
  • GUI ONLY: You'll have to install meslo nerd font without brew. One way is to clone the nerd-fonts repo and run ./install Meslo.

Ubuntu 18.04 server (assumes nvidia and cuda are installed, preferably using a nvidia or pytorch image on a cloudprovider):

####### NOTE: manually perform any `MANUAL` entries: #######
## zsh
sudo apt install zsh
./setup_zsh.sh
cp .zshrc ~/.zshrc

## nvtop
./setup_nvtop.sh

## conda
wget https://repo.anaconda.com/miniconda/Miniconda2-latest-Linux-x86_64.sh
bash Miniconda2-latest-Linux-x86_64.sh

conda create -n fastai python=3.6
## MANUAL: Copy conda setup at the end of .bashrc to .zshrc
## MANUAL: Relog/re-ssh


## fastai
conda activate fastai
# This might take a few mins, you can run the other commands in parallel if you'd like.
conda install pytorch torchvision cudatoolkit=10.0 -c pytorch
pip install jupyter
pip install jupyter_contrib_nbextensions
jupyter contrib nbextension install --user
pip install jupyter_nbextensions_configurator
jupyter nbextensions_configurator enable --user
# MANUAL: when you launch `jupyter notebook`, be sure to enable collapsible headings (and any other extension) in the nbextensions menu

pip install fastai
pip install sklearn