Folder partition

Split your dataset into training and validation folders through terminal, without code implementation.

Installation

You can download folder_partition by cloning the Git repository:

git clone https://github.com/Biscoitinhoo/folder-partition/

Usage

To see the help section, use:

python3 folder_partition.py -h

Arguments:

-f      Path to the dataset. (positional)
-q      Quantity of data to be inserted into training folder. The difference will be inserted into validation. (optional)
-p      Percentage to be inserted into training/validation folder (max 100). (optional)

Examples

All examples take in consideration 1000 data into each folder.

Split data in default mode (80% of the data into training folder, 20% into validation)

python3 folder-partition.py -f /home/biscoitinho/dogs-and-cats
Output: 800 files into training, 200 into validation  

Split data with specific quantity.

python3 folder-partition.py -f /home/biscoitinho/dogs-and-cats -q 600
Output: 600 files into training, 400 into validation.

Split data using percentage

python3 folder-partition.py -f /home/biscoitinho/dogs-and-cats -p 50
Output: 500 files into training, 500 into validation.