Implementation of the table detection and table structure recognition deep learning model described in the paper "ClusterTabNet: Supervised clustering method for table detection and table structure recognition" https://arxiv.org/abs/2402.07502
The requirements are detailed in the requirements.txt file
For sample inference and training, please check out the jupyter notebook: demo.ipynb
Download datasets PubTables-1M, pubtabnet, fintabnet, synthtabnet, icdar2019 and format them using notebooks in the train_data_preparation folder.
To run the evaluation and further training you can call:
CUDA_VISIBLE_DEVICES=0 python train/table_extraction.py --output_dir=OUTPUT_DIRECTORY -t=both --ocr_labels_folder=ocr --learning_rate=0.00001 --is_use_4_points --is_use_image_patches --use_dox_datasets --eval_set='test' --checkpoint_path=model_weights/table_recognition.pth
No known issues
Create an issue in this repository if you find a bug or have questions about the content.
For additional support, ask a question in SAP Community.
If you wish to contribute code, offer fixes or improvements, please send a pull request. Due to legal reasons, contributors will be asked to accept a DCO when they create the first pull request to this project. This happens in an automated fashion during the submission process. SAP uses the standard DCO text of the Linux Foundation.
Copyright (c) 2024 SAP SE or an SAP affiliate company. All rights reserved. This project is licensed under the Apache Software License, version 2.0 except as noted otherwise in the LICENSE file.