Simple tool to split coco annotations (json) into train and test sets.
For spliting new datasets with coco annotation style. To train a computer vision model, we only need the images, annotations, categories attributes. Hence, annotation files of new datasets usually contain only these information but not the info, licenses attributes. However. the original repo assumes these attributes, and thus does not work for these datasets. Also, the original repo pretty-printed the json, increasing the size of the output json files. This fork produces minimal output json files.
cocosplit requires python 3 and a basic set of dependencies:
pip install -r requirements.txt
$ 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
$ python cocosplit.py --having-annotations -s 0.8 /path/to/your/coco_annotations.json train.json test.json
will split coco_annotation.json into train.json and test.json with ratio 80%/20% respectively. It will skip all
images (--having-annotations) without annotations.