Restructure README: cleaner intro, TOC table, common options

- Punchier intro with model count (13 models, 0.9B-8B)
- Quick start example switched to GLM-OCR
- Added collapsible TOC table with all scripts sorted by model size
- Moved common options section up with --help tip
- Added GLM-OCR usage example (OCR, table, formula modes)
- Collapsed detailed per-model docs into expandable section
- Stripped most emoji from headings
- Fixed stale PaddleOCR SOTA claim (GLM-OCR now scores higher)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

README.md

@@ -7,9 +7,9 @@ tags: [uv-script, ocr, vision-language-model, document-processing, hf-jobs]
 
 > Part of [uv-scripts](https://huggingface.co/uv-scripts) - ready-to-run ML tools powered by UV and HuggingFace Jobs.
 
-Ready-to-run OCR scripts that work with `uv run` and HuggingFace Jobs - no setup required!
+13 OCR models from 0.9B to 8B parameters. Pick a model, point at your dataset, get markdown — no setup required.
 
-## 🚀 Quick Start with HuggingFace Jobs
+## 🚀 Quick Start
 
 Run OCR on any dataset without needing your own GPU:
 
@@ -17,25 +17,97 @@ Run OCR on any dataset without needing your own GPU:
 # Quick test with 10 samples
 hf jobs uv run --flavor l4x1 \
     --secrets HF_TOKEN \
-    https://huggingface.co/datasets/uv-scripts/ocr/raw/main/nanonets-ocr.py \
+    https://huggingface.co/datasets/uv-scripts/ocr/raw/main/glm-ocr.py \
     your-input-dataset your-output-dataset \
     --max-samples 10
 ```
 
 That's it! The script will:
 
-- ✅ Process first 10 images from your dataset
-- ✅ Add OCR results as a new `markdown` column
-- ✅ Push the results to a new dataset
-- 📊 View results at: `https://huggingface.co/datasets/[your-output-dataset]`
+- Process first 10 images from your dataset
+- Add OCR results as a new `markdown` column
+- Push the results to a new dataset
+- View results at: `https://huggingface.co/datasets/[your-output-dataset]`
+
+<details><summary>All scripts at a glance (sorted by model size)</summary>
+
+| Script | Model | Size | Backend | Notes |
+|--------|-------|------|---------|-------|
+| `smoldocling-ocr.py` | [SmolDocling](https://huggingface.co/ds4sd/SmolDocling-256M-preview) | 256M | Transformers | DocTags structured output |
+| `glm-ocr.py` | [GLM-OCR](https://huggingface.co/zai-org/GLM-OCR) | 0.9B | vLLM | 94.62% OmniDocBench V1.5 |
+| `paddleocr-vl.py` | [PaddleOCR-VL](https://huggingface.co/PaddlePaddle/PaddleOCR-VL) | 0.9B | Transformers | 4 task modes (ocr/table/formula/chart) |
+| `paddleocr-vl-1.5.py` | [PaddleOCR-VL-1.5](https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.5) | 0.9B | Transformers | 94.5% OmniDocBench, 6 task modes |
+| `lighton-ocr.py` | [LightOnOCR-1B](https://huggingface.co/lightonai/LightOnOCR-1B-1025) | 1B | vLLM | Fast, 3 vocab sizes |
+| `lighton-ocr2.py` | [LightOnOCR-2-1B](https://huggingface.co/lightonai/LightOnOCR-2-1B) | 1B | vLLM | 7× faster than v1, RLVR trained |
+| `hunyuan-ocr.py` | [HunyuanOCR](https://huggingface.co/tencent/HunyuanOCR) | 1B | vLLM | Lightweight VLM |
+| `dots-ocr.py` | [DoTS.ocr](https://huggingface.co/Tencent/DoTS.ocr) | 1.7B | vLLM | 100+ languages |
+| `nanonets-ocr.py` | [Nanonets-OCR-s](https://huggingface.co/nanonets/Nanonets-OCR-s) | 2B | vLLM | LaTeX, tables, forms |
+| `dots-ocr-1.5.py` | [DoTS.ocr-1.5](https://huggingface.co/Tencent/DoTS.ocr-1.5) | 3B | vLLM | Updated multilingual model |
+| `nanonets-ocr2.py` | [Nanonets-OCR2-3B](https://huggingface.co/nanonets/Nanonets-OCR2-s) | 3B | vLLM | Next-gen, Qwen2.5-VL base |
+| `deepseek-ocr-vllm.py` | [DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR) | 4B | vLLM | 5 resolution + 5 prompt modes |
+| `deepseek-ocr.py` | [DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR) | 4B | Transformers | Same model, Transformers backend |
+| `deepseek-ocr2-vllm.py` | [DeepSeek-OCR-2](https://huggingface.co/deepseek-ai/DeepSeek-OCR-2) | 3B | vLLM | Newer, requires nightly vLLM |
+| `olmocr2-vllm.py` | [olmOCR-2-7B](https://huggingface.co/allenai/olmOCR-2-7B-1025-FP8) | 7B | vLLM | 82.4% olmOCR-Bench |
+| `rolm-ocr.py` | [RolmOCR](https://huggingface.co/reducto/RolmOCR) | 7B | vLLM | Qwen2.5-VL based, general-purpose |
+| `numarkdown-ocr.py` | [NuMarkdown-8B](https://huggingface.co/numind/NuMarkdown-8B-Thinking) | 8B | vLLM | Reasoning-based OCR |
+
+</details>
+
+## Common Options
+
+All scripts accept the same core flags. Model-specific defaults (batch size, context length, temperature) are tuned per model based on model card recommendations and can be overridden.
+
+| Option | Description |
+|--------|-------------|
+| `--image-column` | Column containing images (default: `image`) |
+| `--output-column` | Output column name (default: `markdown`) |
+| `--split` | Dataset split (default: `train`) |
+| `--max-samples` | Limit number of samples (useful for testing) |
+| `--private` | Make output dataset private |
+| `--shuffle` | Shuffle dataset before processing |
+| `--seed` | Random seed for shuffling (default: `42`) |
+| `--batch-size` | Images per batch (default varies per model) |
+| `--max-model-len` | Max context length (default varies per model) |
+| `--max-tokens` | Max output tokens (default varies per model) |
+| `--gpu-memory-utilization` | GPU memory fraction (default: `0.8`) |
+| `--config` | Config name for Hub push (for benchmarking) |
+| `--create-pr` | Push as PR instead of direct commit |
+| `--verbose` | Log resolved package versions after run |
+
+Every script supports `--help` to see all available options:
+
+```bash
+uv run glm-ocr.py --help
+```
+
+## Example: GLM-OCR
+
+[GLM-OCR](https://huggingface.co/zai-org/GLM-OCR) (0.9B) scores 94.62% on OmniDocBench V1.5 and supports OCR, formula, and table extraction:
+
+```bash
+# Basic OCR
+hf jobs uv run --flavor l4x1 -s HF_TOKEN \
+    https://huggingface.co/datasets/uv-scripts/ocr/raw/main/glm-ocr.py \
+    my-documents my-ocr-output
+
+# Table extraction
+hf jobs uv run --flavor l4x1 -s HF_TOKEN \
+    https://huggingface.co/datasets/uv-scripts/ocr/raw/main/glm-ocr.py \
+    my-documents my-tables --task table
 
-## 📋 Available Scripts
+# Test on 10 samples first
+hf jobs uv run --flavor l4x1 -s HF_TOKEN \
+    https://huggingface.co/datasets/uv-scripts/ocr/raw/main/glm-ocr.py \
+    my-documents my-test --max-samples 10
+```
 
-### PaddleOCR-VL-1.5 (`paddleocr-vl-1.5.py`) 🏆 SOTA with 6 task modes!
+<details><summary>Detailed per-model documentation</summary>
 
-Ultra-compact SOTA OCR using [PaddlePaddle/PaddleOCR-VL-1.5](https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.5) with 94.5% accuracy:
+### PaddleOCR-VL-1.5 (`paddleocr-vl-1.5.py`) — 6 task modes
 
-- 🏆 **SOTA Performance** - 94.5% on OmniDocBench v1.5
+OCR using [PaddlePaddle/PaddleOCR-VL-1.5](https://huggingface.co/PaddlePaddle/PaddleOCR-VL-1.5) with 94.5% accuracy:
+
+- **94.5% on OmniDocBench v1.5** (0.9B parameters)
 - 🧩 **Ultra-compact** - Only 0.9B parameters
 - 📝 **OCR mode** - General text extraction to markdown
 - 📊 **Table mode** - HTML table recognition
@@ -549,70 +621,6 @@ uv run nanonets-ocr2.py documents ocr-results
 
 ```
 
-## 📁 Works With
-
-Any HuggingFace dataset containing images - documents, forms, receipts, books, handwriting.
-
-## 🎛️ Configuration Options
-
-### Common Options (All Scripts)
-
-| Option                     | Default            | Description                       |
-| -------------------------- | ------------------ | --------------------------------- |
-| `--image-column`           | `image`            | Column containing images          |
-| `--batch-size`             | `32`/`16`\*        | Images processed together         |
-| `--max-model-len`          | `8192`/`16384`\*\* | Max context length                |
-| `--max-tokens`             | `4096`/`8192`\*\*  | Max output tokens                 |
-| `--gpu-memory-utilization` | `0.8`              | GPU memory usage (0.0-1.0)        |
-| `--split`                  | `train`            | Dataset split to process          |
-| `--max-samples`            | None               | Limit samples (for testing)       |
-| `--private`                | False              | Make output dataset private       |
-| `--shuffle`                | False              | Shuffle dataset before processing |
-| `--seed`                   | `42`               | Random seed for shuffling         |
-
-\*RolmOCR and DoTS use batch size 16
-\*\*RolmOCR uses 16384/8192
-
-### Script-Specific Options
-
-**PaddleOCR-VL-1.5**:
-
-- `--task-mode`: Task type - `ocr` (default), `table`, `formula`, `chart`, `spotting`, or `seal`
-- `--output-column`: Override default column name (default: `paddleocr_1.5_[task_mode]`)
-- SOTA 94.5% accuracy on OmniDocBench v1.5
-- Uses transformers backend (single image processing for stability)
-
-**PaddleOCR-VL**:
-
-- `--task-mode`: Task type - `ocr` (default), `table`, `formula`, or `chart`
-- `--no-smart-resize`: Disable adaptive resizing (use original image size)
-- `--output-column`: Override default column name (default: `paddleocr_[task_mode]`)
-- Ultra-compact 0.9B model - fastest initialization!
-
-**GLM-OCR**:
-
-- `--task`: Task type - `ocr` (default), `formula`, or `table`
-- `--repetition-penalty`: Repetition penalty (default: 1.1, from official SDK)
-- Near-greedy sampling by default (temperature=0.01, top_p=0.00001) matching official SDK
-- Requires vLLM nightly + transformers>=5.1.0 (handled automatically)
-
-**DeepSeek-OCR**:
-
-- `--resolution-mode`: Quality level - `tiny`, `small`, `base`, `large`, or `gundam` (default)
-- `--prompt-mode`: Task type - `document` (default), `image`, `free`, `figure`, or `describe`
-- `--prompt`: Custom OCR prompt (overrides prompt-mode)
-- `--base-size`, `--image-size`, `--crop-mode`: Override resolution mode manually
-- ⚠️ **Important for HF Jobs**: Add `-e UV_TORCH_BACKEND=auto` for proper PyTorch installation
-
-**RolmOCR**:
-
-- Output column is auto-generated from model name (e.g., `rolmocr_text`)
-- Use `--output-column` to override the default name
-
-**DoTS.ocr**:
-
-- `--prompt-mode`: Choose `ocr` (default), `layout-all`, or `layout-only`
-- `--custom-prompt`: Override with custom prompt text
-- `--output-column`: Output column name (default: `markdown`)
+</details>
 
-💡 **Performance tip**: Increase batch size for faster processing (e.g., `--batch-size 256` on A100)
+Works with any HuggingFace dataset containing images — documents, forms, receipts, books, handwriting.