A 1.3B OCR model just hit 93.68 on real-world document parsing and there's now an AI judge for OCR errors
英文标题：A 1.3B OCR model just hit 93.68 on real-world document parsing and there's now an AI judge for OCR errors
Most OCR models focus on reading documents. MiniCPM-OCR adds a second piece: a model that checks whether the OCR result is actually trustworthy. OpenBMB just released two open-source models: MiniCPM-OCR 1.0 (1.3B): Scores 93.68 on Real5-OmniDocBench, targeting real-world documents like curved
打开原文相关 2 · 机会 2
中文
Most OCR models focus on reading documents. MiniCPM-OCR adds a second piece: a model that checks whether the OCR result is actually trustworthy. OpenBMB just released two open-source models: MiniCPM-OCR 1.0 (1.3B): Scores 93.68 on Real5-OmniDocBench, targeting real-world documents like curved
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
Most OCR models focus on reading documents. MiniCPM-OCR adds a second piece: a model that checks whether the OCR result is actually trustworthy. OpenBMB just released two open-source models: MiniCPM-OCR 1.0 (1.3B): Scores 93.68 on Real5-OmniDocBench, targeting real-world documents like curved