Split Folders: Python Logic for Automated File Partitioning and Oversampling

Developed by jfilter and updated in January 2026, this repository provides a Python-based utility for partitioning file directories into training, validation, and test subsets. It automates the organization of image files into folder structures required for deep learning workflows.

Use Cases

• Using the oversampling feature to balance class distributions in image folders
• Creating 'train', 'val', and 'test' subdirectories for deep learning model input
• Organizing image files into class-labeled folders for classification tasks

Strengths

• MIT licensed for open redistribution
• Includes oversampling functionality for imbalanced classes
• Updated as of January 2026

Limitations

• Does not include any raw image data
• Requires a Python environment for execution

Provenance

Source

jfilter on GitHub

Freshness

Updated as recently as January 2026.