Simple tool to split coco annotations (json) into train and test sets.

Installation

cocosplit requires python 3 and basic set of dependencies:

pip install -r requirements

Usage for splitting annotation only

$ python cocosplit.py -h
usage: cocosplit.py [-h] -s SPLIT [--having-annotations]
                    coco_annotations train test

Splits COCO annotations file into training and test sets.

positional arguments:
  coco_annotations      Path to COCO annotations file.
  train                 Where to store COCO training annotations
  test                  Where to store COCO test annotations

optional arguments:
  -h, --help            show this help message and exit
  -s SPLIT              A percentage of a split; a number in (0, 1)
  --having-annotations  Ignore all images without annotations. Keep only these
                        with at least one annotation

Running

$ ./split_data.sh <path_to_root_folder_container_data>

will split coco_annotation.json into train.json, val.json and test.json with ratio 80%/12%/8% respectively. It will skip all images (--having-annotations) without annotations. It will also put the images into train, val and test folders.