Upload README.md with huggingface_hub

README.md

@@ -7,25 +7,6 @@ tags:
 - 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: 149862
-  dataset_size: 146150
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
 ---
 
 # Document OCR using DeepSeek-OCR
@@ -37,18 +18,18 @@ This dataset contains markdown-formatted OCR results from images in [Alysonhower
 - **Source Dataset**: [Alysonhower/test](https://huggingface.co/datasets/Alysonhower/test)
 - **Model**: [deepseek-ai/DeepSeek-OCR](https://huggingface.co/deepseek-ai/DeepSeek-OCR)
 - **Number of Samples**: 1
-- **Processing Time**: 2.9 minutes
-- **Processing Date**: 2025-10-22 01:23 UTC
+- **Processing Time**: 1.6 minutes
+- **Processing Date**: 2025-10-23 02:02 UTC
 
 ### Configuration
 
 - **Image Column**: `image`
 - **Output Column**: `markdown`
 - **Dataset Split**: `train`
-- **Resolution Mode**: large
-- **Base Size**: 1280
-- **Image Size**: 1280
-- **Crop Mode**: False
+- **Resolution Mode**: gundam
+- **Base Size**: 1024
+- **Image Size**: 640
+- **Crop Mode**: True
 
 ## Model Information
 
@@ -102,7 +83,7 @@ This dataset was generated using the [uv-scripts/ocr](https://huggingface.co/dat
 uv run https://huggingface.co/datasets/uv-scripts/ocr/raw/main/deepseek-ocr.py \
     Alysonhower/test \
     <output-dataset> \
-    --resolution-mode large \
+    --resolution-mode gundam \
     --image-column image
 ```