FireRed OCR

--- license: apache-2.0 base_model: - Qwen/Qwen3-VL-2B-Instruct pipeline_tag: image-to-text --- <p align="center"> <img src="./assets/logo.png" width="600"/> </p> <p align="center" style="line-height: 1;"> <a href="https://huggingface.co/FireRedTeam" target="_blank"><img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-FireRedTeam-ffc107?color=ffc107&logoColor=white" style="display: inline-block;"/></a> <a href="https://huggingface.co/FireRedTeam/FireRed-OCR" target="_blank"><img alt="Hugging Face Model" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-FireRed--OCR-red" style="display: inline-block;"/></a> <a href="https://huggingface.co/spaces/FireRedTeam/FireRed-OCR" target="_blank"><img alt="Demo" src="https://img.shields.io/badge/%F0%9F%92%BB%20Demo-FireRed--OCR-red" style="display: inline-block;"/></a> </p> <p align="center" style="line-height: 1;"> 🤗 <a href="https://huggingface.co/FireRedTeam/FireRed-OCR">HuggingFace</a> | 🖥️ <a href="https://huggingface.co/spaces/FireRedTeam/FireRed-OCR"> Demo</a> | 📄 <a href="https://arxiv.org/abs/2603.01840">Technical Report</a> | 🐈 <a href="https://github.com/FireRedTeam/FireRed-OCR">GitHub</a> </p> <p align="center"> <img src="./assets/teaser.png" width="800"/> <br> <em>Figure 1: Performance comparison on the OmniDocBench v1.5 benchmark. FireRed-OCR achieves state-of-the-art performance among end-to-end solutions, ranking first with a score above 92%.</em> </p> ## 🔥 FireRed-OCR **FireRed-OCR** is a systematic framework designed to specialize general Large Vision-Language Models (LVLMs) into high-performance, pixel-precise structural document parsing experts. General VLMs frequently suffer from **"Structural Hallucination"** (e.g., disordered rows, invented formulas) when processing complex documents. FireRed-OCR addresses this by shifting the paradigm from "impressionist" text generation to "structural engineering," achieving State-of-the-Art (SOTA) results on authoritative benchmarks like OmniDocBench v1.5. ## ✨ Key Features - **SOTA Performance**: Achieves **92.94%** overall score on OmniDocBench v1.5, significantly outperforming DeepSeek-OCR 2, OCRVerse, and massive general VLMs (e.g., Gemini-3.0 Pro，Qwen3-VL-235B). - **Structural Integrity**: Utilizing **Format-Constrained GRPO** (Group Relative Policy Optimization), the model enforc...

Source context: 1430 downloads · 164 likes · Pipeline image-to-text · Repo FireRedTeam/FireRed-OCR