This repository contains the artifacts for the workshop that CentML team ran at the Vector Institute on Feb 10th, 2023.

You can find workshop slides at here.

First, acquire a GPU interactive session

srun --gres=gpu:1 -c 8 --mem 16G -p t4v2 --pty bash

Clone this repository

git clone https://github.com/CentML/VectorWorkshop.git

Install all the dependencies and launch code server

cd VectorWorkshop
bash setup.sh

If you never used VSCode tunning before, follow the last output from the terminal to authenticate via Github

Open your local vscode instance and install Extenstion: Remote - Tunnels (Again, you might need to authenticate)

Open Command Pallet and type "Remote-Tunnels: Connect to Tunnel, you should see gpuxxx Online"

Click the option, this will transfer you to the actual VSCode instance running on the GPU instance.