README.md

# UBL SOS

SOS retail vision dataset with multiple scene types and classes.

At a Glance

Metric	Value
Archives	1
Total images	1733
Primary format	Roboflow-style YOLO vision export
Delivery format	Zip archives inside each folder

Archive Inventory

Archive	Images	Classes	Project	Source
SOS.zip	1733	24	sos-merged-senmh	dataset

Folder Layout

archive.zip
+-- data.yaml
+-- train/
|   +-- images/
|   +-- labels/
+-- valid/ or val/
+-- test/

Working With the Data

1. Extract the archive you want to inspect.
2. Open data.yaml to confirm class names and split locations.
3. Use the image and label pairs directly for YOLO training, validation, or audit workflows.

Notes

• These archives follow a Roboflow export pattern and keep the original data.yaml metadata inside each zip.
• Several folders contain multiple related datasets; the table above lists each archive separately.