New Hungarian document VQA datasets (HuDocVQA, HuDocVQA-manual, HuCCPDF) reveal far lower accuracy for frontier VLMs than English DocVQA, with finetuning plus OCR data recovering part of the gap.
Enriching word vectors with subword information, 2017
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Synthetic Document Question Answering in Hungarian
New Hungarian document VQA datasets (HuDocVQA, HuDocVQA-manual, HuCCPDF) reveal far lower accuracy for frontier VLMs than English DocVQA, with finetuning plus OCR data recovering part of the gap.