Arabic Handwritten OCR 4bit Qwen2.5 VL 3B v2

--- license: apache-2.0 datasets: - saleh-c4/arabic-ocr-images - aamijar/muharaf-public - johnlockejrr/RASAM - johnlockejrr/KHATT_v1.0_dataset language: - ar base_model: - Qwen/Qwen2.5-VL-3B-Instruct pipeline_tag: image-to-text new_version: sherif1313/Arabic-English-handwritten-OCR-v3 --- <p align="center"> <img src="assets/d93f4651-06cd-4e4c-938e-fae97d6cd60c.png" width="400"/> <p> <p align="center"> 💜 <a href="https://github.com/sherif1313/"><b>Github</b></a>&nbsp&nbsp | &nbsp&nbsp🤗 <a href="https://huggingface.co/sherif1313/Arabic-English-handwritten-OCR-v3">Hugging Face</a>&nbsp&nbsp | &nbsp&nbsp📚 <a href="https://github.com/sherif1313/Arabic-English-handwritten-OCR-v3/tree/main">Cookbooks</a>&nbsp&nbsp <br> 🖥️ <a href="https://huggingface.co/spaces/sherif1313/Arabic-English-handwritten-OCR">Demo</a>&nbsp&nbsp </a> </p> # 🕌 Arabic-handwritten-OCR-4bit-Qwen2.5-VL-3B-v2 🥇 The first open-source Arabic model to achieve 97.2% accuracy in extracting Arabic text from historical books and manuscripts. Outperforms Google Vision and Teseract in the Arabic context. This model is the highest-performing open-source Arabic model ever for handwriting and manuscripts, trained on 65,747 diverse samples, including: Printed texts (from sources such as a font) Handwriting (from a font collection) Historical manuscripts (rare archival documents) # 📊 Performance Scale Value Explanation Evaluation Loss 0.6564, the lowest globally in Arabic OCR Character Error Rate (CER) 4.51%, excellent (less than 5% = high quality) Word Error Rate (WER) ~9%, very good for handwritten texts Estimated Accuracy 97.2%, outperforms commercial models Average Time 0.30-second inference, fast for live applications # 💡 Unique Features ✅ Supports full text (not isolated characters) ✅ Handles old manuscripts (even low quality) ✅ No complex preprocessing required ✅ Fully open source - customizable ✅ Supports full linguistic context for comprehensive text understanding ✅ Noise-resistant and handles low-quality images 🚀 High efficiency: The 4-bit quantum model improves memory usage by about 50% and only 2% less in terms of text extraction compared to the basic model. # 🗂️ Datasets Used 📚 Public Data Aamijar/muharaf-public - Various printed texts KHATT Dataset - Arabic handwriting in various styles Arabic-OCR-images - Comprehensive OCR dataset Rasam (johnlockejrr/RA...