An Empirical Study of Scaling Law for OCR
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914660456857600 |
|---|---|
| author | Rang, Miao Bi, Zhenni Liu, Chuanjian Wang, Yunhe Han, Kai |
| author_facet | Rang, Miao Bi, Zhenni Liu, Chuanjian Wang, Yunhe Han, Kai |
| contents | The laws of model size, data volume, computation and model performance have been extensively studied in the field of Natural Language Processing (NLP). However, the scaling laws in Optical Character Recognition (OCR) have not yet been investigated. To address this, we conducted comprehensive studies that involved examining the correlation between performance and the scale of models, data volume and computation in the field of text recognition.Conclusively, the study demonstrates smooth power laws between performance and model size, as well as training data volume, when other influencing factors are held constant. Additionally, we have constructed a large-scale dataset called REBU-Syn, which comprises 6 million real samples and 18 million synthetic samples. Based on our scaling law and new dataset, we have successfully trained a scene text recognition model, achieving a new state-ofthe-art on 6 common test benchmarks with a top-1 average accuracy of 97.42%. The models and dataset are publicly available at https://github.com/large-ocr-model/large-ocr-model.github.io. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_00028 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | An Empirical Study of Scaling Law for OCR Rang, Miao Bi, Zhenni Liu, Chuanjian Wang, Yunhe Han, Kai Computer Vision and Pattern Recognition The laws of model size, data volume, computation and model performance have been extensively studied in the field of Natural Language Processing (NLP). However, the scaling laws in Optical Character Recognition (OCR) have not yet been investigated. To address this, we conducted comprehensive studies that involved examining the correlation between performance and the scale of models, data volume and computation in the field of text recognition.Conclusively, the study demonstrates smooth power laws between performance and model size, as well as training data volume, when other influencing factors are held constant. Additionally, we have constructed a large-scale dataset called REBU-Syn, which comprises 6 million real samples and 18 million synthetic samples. Based on our scaling law and new dataset, we have successfully trained a scene text recognition model, achieving a new state-ofthe-art on 6 common test benchmarks with a top-1 average accuracy of 97.42%. The models and dataset are publicly available at https://github.com/large-ocr-model/large-ocr-model.github.io. |
| title | An Empirical Study of Scaling Law for OCR |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2401.00028 |