waffle-utils 0.2.3

Waffle Utils 🥛

Waffle Utils

• waffle util tools
• Waffle Data Convention

Install

• python >= 3.9
• pip install waffle_utils

Examples

Create Dataset from coco format

both example below will result same output

Python Code

from waffle_utils.dataset import Dataset
from waffle_utils.dataset.format import Format

url = "https://github.com/snuailab/waffle_utils/raw/main/mnist.zip"

dummy_zip_file = "mnist.zip"
dummy_dataset_name = "mnist"

dummy_extract_dir = "tmp/extract"
dummy_coco_root_dir = "tmp/extract/raw"
dummy_coco_file = "tmp/extract/exports/coco.json"

network.get_file_from_url(url, dummy_zip_file, create_directory=True)
io.unzip(dummy_zip_file, dummy_extract_dir, create_directory=True)

ds = Dataset.from_coco(
    dummy_dataset_name,
    dummy_coco_file,
    dummy_coco_root_dir,
)

ds = Dataset.from_directory(dummy_dataset_name, dummy_data_root_dir)

ds.split_train_val(train_split_ratio=0.8)
ds.export(Format.YOLO_DETECTION)

CLI

wu get_file_from_url --url https://github.com/snuailab/waffle_utils/raw/main/mnist.zip --file-path tmp/mnist.zip
wu unzip --url tmp/mnist.zip --output-dir tmp/extract
wu from_coco --name mnist --coco-file tmp/extract/exports/coco.json --coco-root-dir tmp/extract/raw
wu split_train_val --name mnist --train-split-ratio 0.8
wu export --name mnist --export-format yolo_detection