yunusserhat/

TextOCR-Dataset

	TextOCR Dataset




	
	Version 0.1




	
	Training Set

Word Annotations: 714,770 (272MB) Images: 21,778 (6.6GB)

Word Annotations: 107,802 (39MB) Images: 3,124

Metadata: 1MB Images: 3,232 (926MB)

License: Data is available under CC BY 4.0 license. Important Note: Numbers in the papers should be reported on the v0.1 test set.

Training and validation set images are sourced… See the full description on the dataset page: https://huggingface.co/datasets/yunusserhat/TextOCR-Dataset.