openbmb/MiniCPM-OCR-1.0
AI model · https://huggingface.co/openbmb/MiniCPM-OCR-1.0
First seen by Oriel on .
AI summary
MiniCPM-OCR 1.0 is a 1.3B open-source model for document understanding that parses text, tables, formulas, charts, seals, and structured information from Chinese and English documents.
- It scores 93.68 on Real5-OmniDocBench and 54.7 on OCRBench v2, with leading results reported on OmniDocBench v1.6, Wild-OmniDocBench, and olmOCR-Bench.
- On an NVIDIA GeForce RTX 5090, it processes 1.2 pages per second on OmniDocBench v1.6; DSpark acceleration is reported to improve end-to-end efficiency by 55%.
- The parsing pipeline combines MiniCPM-OCR 1.0 with PP-DocLayoutV3 and the minicpmvocr SDK to recognize page regions and assemble Markdown in reading order.
- Deployment options include vLLM, SGLang, llama.cpp, Ollama, Transformers, and FlagOS, with SDK, API, agent workflows, and fine-tuning support.
- The model uses a SigLIP vision encoder and Qwen3.5 language model, supports Chinese and English, and is released under Apache License 2.0.
#ocr #document understanding #multimodal models #document parsing
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Likes
Total likes reported by the source.
- Hugging Face Models: 442, latest observation ; +264 over observed period since
Downloads (30d)
Rolling 30-day downloads. A change in this value is not new downloads during the observed period.
- Hugging Face Models: 18,359, latest observation ; +18,350 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Models | openbmb/MiniCPM-OCR-1.0 | https://huggingface.co/openbmb/MiniCPM-OCR-1.0 Read discussion |