kiminjo / hil-data-splitter

This repository functions to split the data into a form suitable for HIL.

HIL-data-splitter

This repository is responsible for organizing data in the form of Human in the loop (HIL).

As of September 8, 2023, this repository has been code-implemented to suit the Mobile Phone Defect Segmentation Dataset provided by Kaggle.

If there is a case where HIL needs to be done using other data in the future, I plan to modify the code more generally.

How to Use

1. Download data on kaggle and put it on data directory
   download mobile phone defect data from below URL. https://www.kaggle.com/datasets/girish17019/mobile-phone-defect-segmentation-dataset

2. Run main.py

   python main.py \
       --data data/mobile_data \
       --classes oil scratch stain \
       --save_path results/
   

   change parsers on your case.