Upload README.md with huggingface_hub

README.md

@@ -1,40 +1,55 @@
 ---
 tags:
 - ocr
-- structured-extraction
 - document-processing
-- nuextract3
+- unlimited-ocr
+- baidu
 - markdown
 - uv-script
 - generated
 ---
 
-# NuExtract3 on davanstrien/moh-bench-sample
+# Document OCR using Unlimited-OCR
 
-This dataset contains outputs from [davanstrien/moh-bench-sample](https://huggingface.co/datasets/davanstrien/moh-bench-sample) processed with [NuExtract3](https://huggingface.co/numind/NuExtract3), a 4B vision-language model for document understanding.
+This dataset contains OCR results for [davanstrien/moh-bench-sample](https://huggingface.co/datasets/davanstrien/moh-bench-sample)
+produced by [baidu/Unlimited-OCR](https://huggingface.co/baidu/Unlimited-OCR) with vLLM.
 
 ## Processing Details
 
 - **Source Dataset**: [davanstrien/moh-bench-sample](https://huggingface.co/datasets/davanstrien/moh-bench-sample)
-- **Model**: [numind/NuExtract3](https://huggingface.co/numind/NuExtract3)
-- **Mode**: markdown
+- **Model**: [baidu/Unlimited-OCR](https://huggingface.co/baidu/Unlimited-OCR)
 - **Number of Samples**: 50
-- **Processing Time**: 9.7 min
-- **Processing Date**: 2026-07-08 16:48 UTC
+- **Processing Time**: 6.5 min
+- **Processing Date**: 2026-07-08 16:49 UTC
+- **Output Column**: `markdown`
+- **Split**: `train`
 
-### Configuration
+## Output
 
-- **Image Column**: `image`
-- **Output Column**: `markdown`
-- **Dataset Split**: `train`
-- **Temperature**: 0.2
-- **Thinking Mode**: disabled
+Grounding markup was stripped (`--strip-grounding`); the column holds clean text.
+
+Tables are returned as HTML and equations as LaTeX.
+
+## Usage
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("davanstrien/ocr-bench-moh", split="train")
+print(ds[0]["markdown"])
+```
 
-## Dataset Structure
+## Reproduction
 
-Original columns plus:
-- `markdown`: NuExtract3 output (markdown)
-- `inference_info`: JSON list tracking models applied to this dataset
+Generated with the [uv-scripts/ocr](https://huggingface.co/datasets/uv-scripts/ocr) Unlimited-OCR
+vLLM recipe. Unlimited-OCR needs Baidu's dedicated vLLM image:
 
+```bash
+hf jobs uv run --flavor l4x1 -s HF_TOKEN \
+    --image vllm/vllm-openai:unlimited-ocr --python /usr/bin/python3 \
+    -e PYTHONPATH=/usr/local/lib/python3.12/dist-packages \
+    https://huggingface.co/datasets/uv-scripts/ocr/raw/main/unlimited-ocr-vllm.py \
+    davanstrien/moh-bench-sample <output-dataset>
+```
 
 Generated with [UV Scripts](https://huggingface.co/uv-scripts)