| license: apache-2.0 | |
| # OCRFlux-pubtabnet-single | |
| OCRFlux-pubtabnet-single is a benchmark of 9064 table images and their corresponding ground-truth HTML, which are derived from the public [PubTabNet](https://github.com/ibm-aur-nlp/PubTabNet) benchmark with some format transformations. | |
| This dataset can be used to measure the performance of OCR systems in single-page table parsing. | |
| Quick links: | |
| - 🤗 [Model](https://huggingface.co/ChatDOC/OCRFlux-3B) | |
| - 🛠️ [Code](https://github.com/chatdoc-com/OCRFlux) | |
| ## Data Mix | |
| ## Table 1: Tables breakdown by complexity (whether they contain rowspan or colspan cells) | |
| | Complexity | Number | | |
| |--------|-------------| | |
| | Simple | 4623 | | |
| | Complex | 4441 | | |
| | **Total** | **9064** | | |
| ## Data Format | |
| Each row in the dataset corresponds to a table image and its corresponding ground-truth HTML. | |
| Different from the original PubTabNet dataset, we do not distinguish cells in the table headers and table bodies, which means there are no `<thead>` and `<tbody>` tags, and all `<th>` tags are replaced by `<td>` tags. | |
| ### Features: | |
| ```python | |
| { | |
| 'image_name': string, # Name of the table image | |
| 'type': string, # "simple" or "complex" | |
| 'gt_table': string, # Ground-truth HTML of the table | |
| } | |
| ``` | |
| ## License | |
| This dataset is licensed under Apache-2.0. | |