Skip to content

Latest commit

 

History

History
35 lines (25 loc) · 1.07 KB

README.md

File metadata and controls

35 lines (25 loc) · 1.07 KB

Dev Playground

Simple containerized environment to quickly try random stuff, with a focus on python and deep learning tools.

Also includes some experiments and results.

Features

  • Dockerfile:
    • ubuntu24.04
    • nvidia-cuda12.6
    • cpp tools
    • python tools
    • shell tools
    • xauth/X11
    • miniconda
  • vscode configs (vscode/settings.json and devcontainer.json)
    • gpu support
    • graphical interface support (X11)
    • host camera support
  • Pre-commit hooks
  • Python tool configs (linters, formatters)
  • CI/CD (github actions) for formatting/linting

Explored topics

  • devcontainer: containerized docker environment with dev tools and configs.

  • diffusion: data generation using DDPMs (conditional and unconditional generation).

  • dataset_creator: CLIP based text and image search over an image directory and autolabeling using foundational models.

  • dataset_selection: Selecting of most diverse images of a directory based on similarity algorithm (classical CV).