Buckets:
| task_categories: | |
| - image-to-text | |
| pretty_name: ATR benchmark | |
| size_categories: | |
| - n<1K | |
| language: | |
| - fr | |
| - la | |
| - en | |
| - no | |
| - ar | |
| - zh | |
| - de | |
| - nl | |
| tags: | |
| - atr | |
| - htr | |
| - ocr | |
| dataset_info: | |
| features: | |
| - name: dataset | |
| dtype: string | |
| - name: image | |
| dtype: image | |
| - name: text | |
| dtype: string | |
| - name: level | |
| dtype: string | |
| splits: | |
| - name: test | |
| num_bytes: 180132193.0 | |
| num_examples: 133 | |
| download_size: 178873479 | |
| dataset_size: 180132193.0 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: test | |
| path: data/test-* | |
| # ATR benchmark - Page/paragraph level | |
| ## Dataset Description | |
| - **Homepage:** [ATR benchmark](https://huggingface.co/datasets/Teklia/ATR-benchmark) | |
| - **Point of Contact:** [TEKLIA](https://teklia.com) | |
| ## Dataset Summary | |
| The ATR benchmark dataset is a multilingual dataset that includes 463 document images, at page or paragraph level. This dataset has been designed to test ATR models and combines data from several public datasets: | |
| - [BnL Historical Newspapers](https://data.bnl.lu/data/historical-newspapers/) | |
| - [CASIA-HWDB2](https://nlpr.ia.ac.cn/databases/handwriting/Offline_database.html) | |
| - [Churro](https://huggingface.co/datasets/stanford-oval/churro-dataset) | |
| - [DIY History - Social Justice](http://diyhistory.lib.uiowa.edu/) | |
| - [DAI-CRETDHI](https://dai-cretdhi.univ-lr.fr/) | |
| - [Esposalles](https://dag.cvc.uab.es/dataset/the-esposalles-database/) | |
| - FINLAM - Historical Newspapers | |
| - [Horae - Books of hours](https://github.com/oriflamms/HORAE) | |
| - [IAM](https://fki.tic.heia-fr.ch/databases/iam-handwriting-database) | |
| - [NDLOCR](https://lab.ndl.go.jp/data_set/r4_kotenocr_en/) | |
| - [NorHand v3](https://zenodo.org/records/10255840) | |
| - [OpenITI](https://openiti.org/projects/OpenITI%20Corpus.html) | |
| - [Marius PELLET](https://europeana.transcribathon.eu/documents/story/?story=121795) | |
| - [QARI](https://huggingface.co/datasets/NAMAA-Space/QariOCR-v0.3-markdown-mixed-dataset) | |
| - [RASM](http://www.primaresearch.org/RASM2019/) | |
| - [READ-2016](https://zenodo.org/records/218236) | |
| - [RIMES](https://teklia.com/research/rimes-database/) | |
| - [ScribbleLens](https://openslr.org/84/) | |
| Images are in their original size. | |
| ### Split | |
| | dataset | images | language | | |
| | ------------------------------ | -----: | ---------------- | | |
| | BnL Historical Newspapers | 3 | German | | |
| | CASIA-HWDB2 | 10 | Chinese | | |
| | Churro | 290 | Multi-lingual | | |
| | DAI-CRETDHI | 10 | French | | |
| | DIY History - Social Justice | 20 | English | | |
| | Esposalles | 10 | Catalan | | |
| | FINLAM - Historical Newspapers | 10 | English / French | | |
| | Horae - Books of hours | 10 | Latin | | |
| | IAM | 10 | English | | |
| | NDLOCR | 10 | Japanese | | |
| | NorHand v3 | 10 | Norwegian | | |
| | OpenITI | 10 | Arabic | | |
| | Marius PELLET | 10 | French | | |
| | QARI | 10 | Arabic | | |
| | RASM | 10 | Arabic | | |
| | READ-2016 | 10 | German | | |
| | RIMES | 10 | French | | |
| | ScribbleLens | 10 | Dutch | | |
| ## Dataset Structure | |
| ### Data Instances | |
| ``` | |
| { | |
| 'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size={img size} at 0x1A800E8E190, | |
| 'text': '{transcription}' | |
| } | |
| ``` | |
| ### Data Fields | |
| - `image`: a PIL.Image.Image object containing the image. Note that when accessing the image column (using dataset[0]["image"]), the image file is automatically decoded. Decoding of a large number of image files might take a significant amount of time. Thus it is important to first query the sample index before the "image" column, i.e. dataset[0]["image"] should always be preferred over dataset["image"][0]. | |
| - `dataset`: the name of the original dataset. | |
| - `text`: the label transcription of the image. | |
| - `level`: full document page or a single paragraph. | |
Xet Storage Details
- Size:
- 4.18 kB
- Xet hash:
- 7de4f23544f5799ac5b600da8546b6a2631d3390093600ffeced1b27bdd8aed5
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.