Upload README.md with huggingface_hub

README.md

@@ -2,67 +2,53 @@
 tags:
 - ocr
 - document-processing
-- nanonets
-- nanonets-ocr2
+- deepseek
+- deepseek-ocr
 - markdown
 - uv-script
 - generated
-dataset_info:
-  features:
-  - name: image
-    dtype: image
-  - name: markdown
-    dtype: string
-  - name: inference_info
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 146150
-    num_examples: 1
-  download_size: 149864
-  dataset_size: 146150
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
 ---
 
-# Document OCR using Nanonets-OCR2-3B
+# Document OCR using DeepSeek-OCR
 
-This dataset contains markdown-formatted OCR results from images in [Alysonhower/test](https://huggingface.co/datasets/Alysonhower/test) using Nanonets-OCR2-3B.
+This dataset contains markdown-formatted OCR results from images in [Alysonhower/test](https://huggingface.co/datasets/Alysonhower/test) using DeepSeek-OCR.
 
 ## Processing Details
 
 - **Source Dataset**: [Alysonhower/test](https://huggingface.co/datasets/Alysonhower/test)
-- **Model**: [nanonets/Nanonets-OCR2-3B](https://huggingface.co/nanonets/Nanonets-OCR2-3B)
-- **Model Size**: 3.75B parameters
+- **Model**: [deepseek-ai/DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR)
 - **Number of Samples**: 1
-- **Processing Time**: 9.1 minutes
-- **Processing Date**: 2025-10-14 22:11 UTC
+- **Processing Time**: 1.5 minutes
+- **Processing Date**: 2025-10-22 01:07 UTC
 
 ### Configuration
 
 - **Image Column**: `image`
 - **Output Column**: `markdown`
 - **Dataset Split**: `train`
-- **Batch Size**: 16
-- **Max Model Length**: 15,000 tokens
-- **Max Output Tokens**: 15,000
-- **GPU Memory Utilization**: 80.0%
+- **Resolution Mode**: gundam
+- **Base Size**: 1024
+- **Image Size**: 640
+- **Crop Mode**: True
 
 ## Model Information
 
-Nanonets-OCR2-3B is a state-of-the-art document OCR model that excels at:
+DeepSeek-OCR is a state-of-the-art document OCR model that excels at:
 - 📐 **LaTeX equations** - Mathematical formulas preserved in LaTeX format
-- 📊 **Tables** - Extracted and formatted as HTML
+- 📊 **Tables** - Extracted and formatted as HTML/markdown
 - 📝 **Document structure** - Headers, lists, and formatting maintained
-- 🖼️ **Images** - Captions and descriptions included in `<img>` tags
-- ☑️ **Forms** - Checkboxes rendered as ☐/☑
-- 🔖 **Watermarks** - Wrapped in `<watermark>` tags
-- 📄 **Page numbers** - Wrapped in `<page_number>` tags
+- 🖼️ **Image grounding** - Spatial layout and bounding box information
+- 🔍 **Complex layouts** - Multi-column and hierarchical structures
 - 🌍 **Multilingual** - Supports multiple languages
 
+### Resolution Modes
+
+- **Tiny** (512×512): Fast processing, 64 vision tokens
+- **Small** (640×640): Balanced speed/quality, 100 vision tokens
+- **Base** (1024×1024): High quality, 256 vision tokens
+- **Large** (1280×1280): Maximum quality, 400 vision tokens
+- **Gundam** (dynamic): Adaptive multi-tile processing for large documents
+
 ## Dataset Structure
 
 The dataset contains all original columns plus:
@@ -91,23 +77,22 @@ for info in inference_info:
 
 ## Reproduction
 
-This dataset was generated using the [uv-scripts/ocr](https://huggingface.co/datasets/uv-scripts/ocr) Nanonets OCR2 script:
+This dataset was generated using the [uv-scripts/ocr](https://huggingface.co/datasets/uv-scripts/ocr) DeepSeek OCR script:
 
 ```bash
-uv run https://huggingface.co/datasets/Alysonhower/scripts/resolve/main/nanonets-ocr2.pyy \
+uv run https://huggingface.co/datasets/uv-scripts/ocr/raw/main/deepseek-ocr.py \
     Alysonhower/test \
     <output-dataset> \
-    --model nanonets/Nanonets-OCR2-3B \
-    --image-column image \
-    --batch-size 16 \
-    --max-model-len 15000 \
-    --max-tokens 15000 \
-    --gpu-memory-utilization 0.8
+    --resolution-mode gundam \
+    --image-column image
 ```
 
 ## Performance
 
 - **Processing Speed**: ~0.0 images/second
-- **GPU Configuration**: vLLM with 80% GPU memory utilization
+- **Processing Method**: Sequential (Transformers API, no batching)
+
+Note: This uses the official Transformers implementation. For faster batch processing,
+consider using the vLLM version once DeepSeek-OCR is officially supported by vLLM.
 
 Generated with 🤖 [UV Scripts](https://huggingface.co/uv-scripts)