{"slug": "arabic-handwritten-vlm", "title": "Arabic handwritten VLM", "summary": "A user is seeking recommendations for the most accurate vision-language model (VLM) for extracting text from handwritten Arabic forms such as applications, questionnaires, and administrative documents. The request asks for benchmark results, model recommendations, and fine-tuning advice on models including Qwen-VL, Florence, PaliGemma, and Arabic-specific OCR/VLM systems, with a preference for offline or self-hosted inference. The user specifically wants high accuracy in Arabic handwritten text recognition, structured form understanding of fields, tables, and checkboxes, and mixed printed and handwritten Arabic content.", "body_md": "Hi everyone,\n\nBased on your experience, which trained VLM performs best for extracting text from handwritten Arabic forms (e.g., applications, questionnaires, administrative documents)?\n\nI’m mainly looking for the highest possible accuracy in:\n\n- Arabic handwritten text recognition\n- Structured form understanding (fields, tables, checkboxes)\n- Mixed printed and handwritten Arabic content\n- Ideally offline/self-hosted inference\n\nHave you tested any models such as Qwen-VL, Florence, PaliGemma, or Arabic-specific OCR/VLM models on this type of document? If possible, I’d appreciate benchmark results, model recommendations, and fine-tuning advice.\n\nThank you!", "url": "https://wpnews.pro/news/arabic-handwritten-vlm", "canonical_source": "https://discuss.huggingface.co/t/arabic-handwritten-vlm/180583#post_1", "published_at": "2026-09-18 09:17:44+00:00", "updated_at": "2026-09-18 09:25:24.300185+00:00", "lang": "en", "topics": ["computer-vision", "natural-language-processing", "ai-tools", "ai-research"], "entities": ["Qwen-VL", "Florence", "PaliGemma"], "alternates": {"html": "https://wpnews.pro/news/arabic-handwritten-vlm", "markdown": "https://wpnews.pro/news/arabic-handwritten-vlm.md", "text": "https://wpnews.pro/news/arabic-handwritten-vlm.txt", "jsonld": "https://wpnews.pro/news/arabic-handwritten-vlm.jsonld"}}