fastai / fastsetup

Setup all the things
Apache License 2.0
228 stars 87 forks source link

fastsetup

Setup all the things

First, do basic ubuntu configuration, such as updating packages, and turning on auto-updates:

sudo apt update && sudo apt -y install git
git clone https://github.com/fastai/fastsetup.git
cd fastsetup
sudo ./ubuntu-initial.sh
# wait a couple of minutes for reboot, then ssh back in
# If you're using WSL (Windows) use `sudo ./ubuntu-wsl.sh` instead of the above line

Then, optionally, set up dotfiles:

source dotfiles.sh

...and set up conda:

source setup-conda.sh
. ~/.bashrc
conda install -yq mamba

To set up email:

sudo ./opensmtpd-install.sh

To test email, create a text file msg containing a message to send, then send it with:

cat msg |  mail -r "x@$(hostname -d)" -s 'subject' EMAIL_ADDR

Replace EMAIL_ADDR with an address to send to. You can get a useful testing address from mail-tester.

NVIDIA drivers

To install NVIDIA drivers, if required:

ubuntu-drivers devices
sudo apt-fast install -y nvidia-XXX{-server}
sudo modprobe nvidia
nvidia-smi

WSL-only:

Install the latest NVIDIA driver on your Windows PC running WSL Then do a special install of cuda

sudo apt-key del 7fa2af80
wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda-repo-wsl-ubuntu-11-8-local_11.8.0-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-11-8-local_11.8.0-1_amd64.deb
sudo cp /var/cuda-repo-wsl-ubuntu-11-8-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-fast install -y cuda
nvidia-smi

(WSL-only): Don't worry if nvidia-smi reports "Internal Error" under the "Processes" heading.

If it's working, you should still see part of your GPU name, and how much memory is available in the first heading.