Small Program for Labeling YOLO (prog.pgs.sh)

A lightweight, offline labeling tool for YOLO has been released (version 0.0.1). Designed for users frustrated by web-only labelers, the program runs locally and organizes datasets into a simple filesystem layout: an images/ directory with subfolders of pictures, a parallel labels/ tree to store per-image .txt YOLO annotations, and a labels mapping file (etykiety.txt) with a sample provided. The UI lists files with color-coded status—green for images that already have labels and red for unlabeled—making it easy to see labeling progress at a glance. Thumbnail samples are drawn from the COCO dataset. Key technical notes: label files follow YOLO .txt format and must be placed in labels/ subfolders mirroring images/. You must provide your own labels mapping in etykiety.txt, or edit the sample. The app is distributed as an AppImage split across three tar archives; download the three parts and recombine/extract with: tar -x -M --file=dysk1.tar --file=dysk2.tar --file=dysk3.tar then mark the resulting .AppImage executable (e.g., chmod +x). Its simplicity, offline operation, and filesystem-based organization make it practical for privacy-sensitive or low-connectivity labeling workflows and quick dataset curation for YOLO training.