datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
docvqa-single-page-questions
Dataset Card for DocVQA Dataset
Dataset Summary
DocVQA dataset is a document dataset introduced in Mathew et al. (2021) consisting of 50,000 questions defined on 12,000+ document images.
Please visit the challenge page (https://rrc.cvc.uab.es/?ch=17) and paper (https://arxiv.org/abs/2007.00398) for further information.
Usage
This dataset can be used with current releases of Hugging Face datasets library.
Here is an example using a custom collator to bundle… See the full description on the dataset page: https://huggingface.co/datasets/pixparse/docvqa-single-page-questions.docvqa-single-page-questions-answer-ocr
DocVQA with Answer Localization
This dataset provides answer-localization annotations produced by our pipeline on top of the DocVQA dataset.
Usage
from datasets import load_dataset
# Load the dataset with answer OCR annotations
ds = load_dataset("indrehus/docvqa-single-page-questions-answer-ocr", split="validation")
# Get a single sample
sample = ds[0]
# Available fields in each sample:
print("Image:", sample["image"]) # PIL.Image
print("Question:"… See the full description on the dataset page: https://huggingface.co/datasets/indrehus/docvqa-single-page-questions-answer-ocr.docvqa-single-page-questions-answer-ocr-colSmol500M-q-priors
DocVQA Evidence Heatmaps (colsmol-500M)
This dataset contains question–evidence aligned heatmaps produced by our pipeline using colsmol-500M. It is intended as an auxiliary artifact to accompany our main dataset:
Main dataset (DocVQA with Answer Localization): https://huggingface.co/datasets/indrehus/docvqa-single-page-questions-answer-ocr
Usage
from datasets import load_dataset
# Load the dataset with ColSmol-500M question priors
ds = load_dataset(… See the full description on the dataset page: https://huggingface.co/datasets/indrehus/docvqa-single-page-questions-answer-ocr-colSmol500M-q-priors.