Upload README.md with huggingface_hub

README.md

@@ -7,45 +7,6 @@ tags:
 - markdown
 - uv-script
 - generated
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: image
-    dtype: image
-  - name: drawer_id
-    dtype: string
-  - name: card_number
-    dtype: int64
-  - name: filename
-    dtype: string
-  - name: text
-    dtype: string
-  - name: has_ocr
-    dtype: bool
-  - name: source
-    dtype: string
-  - name: source_url
-    dtype: string
-  - name: ia_collection
-    dtype: string
-  - name: markdown
-    dtype: string
-  - name: deepseek_markdown
-    dtype: string
-  - name: deepseek_markdown_free
-    dtype: string
-  - name: inference_info
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 33465634
-    num_examples: 500
-  download_size: 31332799
-  dataset_size: 33465634
 ---
 
 # Document OCR using DeepSeek-OCR-2
@@ -57,8 +18,8 @@ This dataset contains markdown-formatted OCR results from images in [davanstrien
 - **Source Dataset**: [davanstrien/bpl-card-catalog-glm-ocr-bench](https://huggingface.co/datasets/davanstrien/bpl-card-catalog-glm-ocr-bench)
 - **Model**: [deepseek-ai/DeepSeek-OCR-2](https://huggingface.co/deepseek-ai/DeepSeek-OCR-2)
 - **Number of Samples**: 500
-- **Processing Time**: 12.1 min
-- **Processing Date**: 2026-02-14 17:11 UTC
+- **Processing Time**: 7.0 min
+- **Processing Date**: 2026-02-14 17:26 UTC
 
 ### Configuration
 
@@ -124,7 +85,7 @@ uv run https://huggingface.co/datasets/uv-scripts/ocr/raw/main/deepseek-ocr2-vll
 
 ## Performance
 
-- **Processing Speed**: ~0.7 images/second
+- **Processing Speed**: ~1.2 images/second
 - **Processing Method**: Batch processing with vLLM (2-3x speedup over sequential)
 
 Generated with [UV Scripts](https://huggingface.co/uv-scripts)