IshaanKaur/

dataset_152150257_document_ocr_image_depth

	dataset_152150257_document_ocr_image_depth.py




	


	Dataset Summary

A document ocr dataset with image depth modality, stored in parquet format.

Preprocessing: aggressive Augmentation: autoaugment

Split strategy: stratified 90 10 Sampling: balanced

Quality filtering: adaptive Labeling: weak supervision