Nayana-cognitivelab/Nayana-OCRBench-in-0.1k-v2-arxiv dataset hosted on Hugging Face and contributed by the HF Datasets community

OCRBench is a comprehensive evaluation benchmark designed to assess the OCR capabilities of Large Multimodal Models. It comprises five components: Text Recognition, SceneText-Centric VQA, Document-Oriented VQA, Key Information Extraction, and Handwritten Mathematical Expression Recognition. The benchmark includes 1000 question-answer pairs, and all the answers undergo manual verification and correction to ensure a more precise evaluation.

Nayana-cognitivelab/Nayana-OCRBench-final dataset hosted on Hugging Face and contributed by the HF Datasets community

Nayana-cognitivelab/Nayana-OCRBench-in-0.6k-v1-arxiv dataset hosted on Hugging Face and contributed by the HF Datasets community

OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning

https://github.com/Yuliang-Liu/MultimodalOCR https://arxiv.org/abs/2501.00321

from datasets import load_dataset
repo_dir = 'morpheushoc/OCRBenchv2'

# load all samples
dataset = load_dataset(repo_dir, split='test') # 10k samples

# load a subset (EN/CN samples)
dataset = load_dataset(repo_dir, 'EN', split='test') # 7.4k samples 
dataset =… See the full description on the dataset page: https://huggingface.co/datasets/morpheushoc/OCRBenchv2.

ahmedheakl/i8n-ocr-bench dataset hosted on Hugging Face and contributed by the HF Datasets community