Ubuntu Use Case

Use an Ubuntu template and SSH to create a remote environment. Then, SSH into it, install the CUDA driver and PyTorch.

Step 1: Create a GPU Container using Ubuntu template

• Choose Ubuntu template

• Select SSH Terminal Access

Step 2: Connect your container via SSH by using VSCode

To connect via SSH, copy the command and paste it into your terminal. You can use any terminal application, such as Command Prompt, PowerShell, or the integrated terminal in VS Code.

See our detailed guide here for a full walkthrough.

Step 3: Setting Up Your Environment

1. Install python3

apt update && apt install -y wget gnupg2 curl software-properties-common 
apt install -y python3 python3-pip python3-venv 

1. Install Pytorch

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124

Check if pytorch is installed successfully

python 3 
import torch 

1. Run a simple matrix algorithm

apt install vim 
vim matrix_mulplication.py 
python3 matrix_mulplication.py