--- license: apache-2.0 language: - en - zh pipeline_tag: image-text-to-text tags: - minicpm-ocr - ocr - correctness-decision - ocr-verification - multimodal - vision-language - jev ---

[GitHub](https://github.com/OpenBMB/MiniCPM-OCR) · [Deploy](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/minicpm_ocr_jev_tutorial_en.md)

**MiniCPM-OCR Jev** is a **2B OCR correctness decision model** for text, formula, and table regions. Given a region image and a candidate OCR result, it returns a **score from 0 to 1** — higher means the candidate is more likely correct. Use it to auto-accept reliable regions, route the rest to human review, or pick the best result among several OCR models. - 🛡️ **A small model that beats larger ones**: at only 2B parameters, decision AUROC (×100) on [JevOCRBench](https://huggingface.co/datasets/openbmb/JevOCRBench) is **93.6**, above JEV-35B-A3B (90.1) and close to Qwen3.5-397B-A17B (93.8). - 🧩 **Stable generalization**: on parses from MiniCPM-OCR 1.0, GLM-OCR, PaddleOCR-VL-1.6, and Xiaomi-OCR-0, AUROC (×100) stays in **91.9–93.4**. ### Architecture Trained from **Qwen3.5-2B**. The vision encoder and language backbone are retained, and the language-model output head `lm_head` is replaced with a **binary classification head** that learns whether a candidate is correct. #### Inputs and outputs | Item | Description | | :--- | :--- | | Assessment unit | One text / formula / table region plus its candidate OCR result | | Original region image | The crop from the document | | Candidate content | OCR text and an image rendered from that text | | Output | Score in 0–1; higher = more likely correct | Internally, the model uses a matched original region image, candidate text, and rendered image. When calling the cloud API, the Docker image, or the SDK service with a built-in renderer, provide the original region image, candidate text, and the correct region category (text, formula, or table); the service handles rendering. The SDK service without a renderer requires the caller to supply the rendered image. See the [deployment guide](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/minicpm_ocr_jev_tutorial_en.md) for request fields and image encoding formats. ### Evaluation Results The table below reports overall and per-category AUROC (×100) on JevOCRBench. Higher values indicate better discrimination between correct and incorrect parsing results.

Overall, text, formula, and table AUROC on JevOCRBench, highlighting MiniCPM-OCR Jev.

### Discrimination across OCR models The cross-parser experiment uses the same 200 samples randomly drawn from JevOCRBench. Four OCR models parse the samples independently; each result is manually labelled correct or incorrect and then scored by MiniCPM-OCR Jev. | OCR model | MiniCPM-OCR 1.0 | GLM-OCR | PaddleOCR-VL-1.6 | Xiaomi-OCR-0 | | :--- | ---: | ---: | ---: | ---: | | AUROC (×100) | 92.7 | 91.9 | 93.4 | 93.1 | In this experiment, AUROC (×100) ranges from 91.9 to 93.4, with a maximum difference of 1.5 points. These results describe performance on the samples and parser combinations above. Validate separately on your target data when using other parsers or new document types. ### Automatic screening to reduce human review Score the parsed result for each region and choose a threshold using samples from your application. Results above the threshold proceed automatically; the rest go to human review. A higher threshold typically reduces the number of results that pass automatically, so monitor both the accuracy of automatically accepted results and the volume of human review.

Accuracy versus review-free rate for legal documents, financial receipts, and complex documents.

### Selecting across models to improve parsing quality Use the same layout segmentation to have several OCR models parse each region. MiniCPM-OCR Jev scores the candidates for the same region, selects the highest-scoring candidate, and combines the selected results in reading order. If all candidates fall below the application threshold, the region can still be sent for human review.

Parsing-quality gains from MiniCPM-OCR Jev scoring for MiniCPM-OCR 1.0, Xiaomi OCR, and PaddleOCR-VL 1.6.

Candidates for the same region must share the crop and content boundaries. Jev selects among existing candidates, so quality is bounded by their coverage and correctness. ### Examples The figure below compares the original image with candidates 1 and 2, showing how MiniCPM-OCR Jev scores text, formula, and table candidates.

Three examples for text, formulas, and tables: original, candidate 1 and candidate 2, with MiniCPM-OCR Jev scores.

## Quick start | Deployment | Guide | | :--- | :--- | | Cloud API | [Single-candidate scoring, best-of-N, batched best-of-N](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/minicpm_ocr_jev_api_en.md) (stub — being written) | | Docker / local SDK | [Environment, rendering options, service startup](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/minicpm_ocr_jev_tutorial_en.md) | | vLLM backend | [Service configuration with and without a renderer](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/vllm_deploy_jev_en.md) | | SGLang | [Deployment guide](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/sglang_deploy_jev_en.md) | | llama.cpp | [GGUF deployment guide](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/llamacpp_deploy_jev_en.md) | | Ollama | [Deployment guide](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/ollama_deploy_jev_en.md) | | FlagOS (multi-chip) | [Deployment guide](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/flagos_deploy_en.md) | When a service has a built-in renderer (cloud API, Docker, SDK), check `verdict` before using the score — only `scored` means model scoring actually ran. Fixed values returned for precheck, rendering, or forward-pass failures must not be read as low model scores. ### Fine-tuning [ms-swift guide](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/finetune_jev_en.md): data format, binary classification head configuration, training commands, and JevOCRBench evaluation. ## Related checkpoints | Format | Hugging Face | ModelScope | | :--- | :--- | :--- | | BF16 (original) **(this repo)** | [🤗](https://huggingface.co/openbmb/MiniCPM-OCR-Jev) | [ModelScope](https://modelscope.cn/models/OpenBMB/MiniCPM-OCR-Jev) | | GGUF (llama.cpp) | [🤗](https://huggingface.co/openbmb/MiniCPM-OCR-Jev-gguf) | [ModelScope](https://modelscope.cn/models/OpenBMB/MiniCPM-OCR-Jev-gguf) | | AWQ W4A16 | [🤗](https://huggingface.co/openbmb/MiniCPM-OCR-Jev-AWQ) | [ModelScope](https://modelscope.cn/models/OpenBMB/MiniCPM-OCR-Jev-AWQ) | | GPTQ W4A16 | [🤗](https://huggingface.co/openbmb/MiniCPM-OCR-Jev-GPTQ) | [ModelScope](https://modelscope.cn/models/OpenBMB/MiniCPM-OCR-Jev-GPTQ) | Jev works on candidates from any parser. The companion parser is [MiniCPM-OCR 1.0](https://huggingface.co/openbmb/MiniCPM-OCR-1.0). ## FAQ Installation, model downloads, GPU memory, input formats, and unexpected outputs: [FAQ](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/docs/faq_en.md). ## License Model weights, code, and documentation are released under the [Apache License 2.0](https://github.com/OpenBMB/MiniCPM-OCR/blob/main/LICENSE).