Python Boilerplate with Pytorch GPU support

📙 Introduction

• PyTorch GPU boilerplate for deep learning projects.
• Also includes Ruff, Pytest, Pre-commit, Poetry set-up for a development environment.
• [Branch: devcontainer] Provides a devcontainer set-up for Visual Studio Code Remote Development and GitHub Codespaces.

🚀 Quick Start

1. Open Visual Studio Code with Remote Development

• Install Visual Studio Code and Remote Development Extension Pack.
• Connect to a (Case 1) Remote-SSH server or create a new (Case 2) Codespaces with this repository.
  1. (Recommend) How to Connect Remote-SSH
  2. Install GitHub Codespaces Extension and Create a new Codespaces with this repository. important! Check you have access to Codespaces plan. (free accounts does not support GPU machines.)
• ℹ️ Be aware that this devcontainer is set up for a GPU environment. you may need a gpu-enabled remote server or Codespaces.

2. Create your repository using template with the devcontainer branch

• Click the Use this template button on the top-right of this repository.
  • How to use a repository as a template
• Click the Create repository button.

3. Open the repository in Visual Studio Code

• Open the repository in Remote Server or Codespaces.

• Click the Reopen in Container button to open the repository in the devcontainer.

  • You may need to install Docker if you are using a local machine.
  • check devcontainer.json for more information about the devcontainer settings.
    • container environment

      key	description	default
      TZ	Timezone	Asia/Seoul
      LANG	Locale	ko_KR.UTF-8
      LC_MESSAGES	Locale (POSIX for command output as eng)	POSIX
      NVIDIA_VISIBLE_DEVICES	GPU device id (0, 1, 2, ...)	all
      NVIDIA_DRIVER_CAPABILITIES	GPU driver capabilities	compute,utility
      SKIP_PACKAGES_UPDATE	Skip package update (poetry, npm, ...)	false
      POETRY_WITHOUT_GROUPS	Skip poetry groups installation

• A new container will be created and the repository will be opened in the new container. (It may take a few minutes to build the container for the first time.)

• (Optional) Click the Reopen in Workspace button to open the repository in the workspace.

  • Additional extension settings will be applied to the workspace.

4. Install additional dependencies

• Check Poetry Managing Dependencies for more information.
• Check Node Package Manager for more information.

5. Start your project

• Start coding your project with the boilerplate.
• Other useful commands are listed in the .config/scripts/bin directory.
  • update-script : Update dependencies (git submodule, poetry, npm, pre-commit)
  • cache-clean : Clean cache files (__pycache__, .pytest_cache, .coverage, ...)

Done! 🎉

---

📔 Additional Information

Commit Signinig with SSH (Verified Commit)

• How to sign commits

note You don't need to set up git on the devcontainer. The git configuration on the local machine will be reflected to the devcontainer automatically.

1. Generate a new SSH key and add it to your GitHub account.

   # Creates a new ssh key, using the provided email as a label
   ssh-keygen -t ed25519 -C "your@github_email"
   
   # Start the ssh-agent in the background (if not already running)
   eval "$(ssh-agent -s)"
   
   # Add your SSH private key to the ssh-agent
   ssh-add PATH/TO/.SSH/PRIVATE_KEY   # e.g. ~/.ssh/id_ed25519

2. Add the SSH key to your GitHub account.

   • Adding a new SSH key to your GitHub account

3. Set global git configuration for signing.

   git config --global gpg.format ssh
   git config --global gpg.ssh.defaultKeyCommand "ssh-add -L"

4. Add following line to your settings.json in Visual Studio Code.

   "git.enableCommitSigning": true

   • Or you can enable it in the git configuration.

     Signing commits

     git config commit.gpgSign true

5. Set your user name and email address to your GitHub username and email.

   📝 Note
   Set name and email for the repository(--local), not globally(--global).

   git config user.name "your_github_username"
   git config user.email "your@github_email"