README.md

Data format example

This implementation requires the input data in the following format:

Training/Validation data

1. train_samples_list.csv file: multi-line with format: index,document_type,file_name.
2. boxes_and_transcripts folder: file_name.tsv files.
   • every file_name.tsv file has multi-line with format: index,box_coordinates (clockwise 8 values), transcripts,box_entity_types .
3. images folder: file_name.jpg files.
4. entities folder (optional) : file_name.txt files.
   • every file_name.txt file contains a json format string, providing the exactly label value of every entity.
   • if iob_tagging_type is set to box_level, this folder will not be used, then box_entity_types in file_name.tsv file of boxes_and_transcripts folder will be used as label of entity. otherwise, it must be provided.

Testing data

1. boxes_and_transcripts folder: file_name.tsv files
   • every file_name.tsv file has multi-line with format: index,box_coordinates (clockwise 8 values), transcripts.
2. images folder: file_name.jpg files.