Fine-tuning Qwen2-VL on synthetic Arabic data yields QARI v0.2 with CER 0.061 and WER 0.160 on the authors' private test set, but public SARD results show Mistral OCR is more accurate.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
QARI-OCR: High-Fidelity Arabic Text Recognition through Multimodal Large Language Model Adaptation
Fine-tuning Qwen2-VL on synthetic Arabic data yields QARI v0.2 with CER 0.061 and WER 0.160 on the authors' private test set, but public SARD results show Mistral OCR is more accurate.