Datasets:

davanstrien
/

ocr-affordances-smoldocling

page_id stringclasses 8 values	page_type stringclasses 1 value	page_number stringclasses 6 values	ocr_text stringclasses 7 values	markdown stringclasses 8 values	inference_info stringclasses 1 value
brit_00000	Normal			0>0>500>500><remote_sensing>	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_00389	Normal	336	PlateXXHI.	11>41>474>469> 373>478>471>490>	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_00778	Normal	647	B O Themp, tranfplanted them at a great diftance from anymales of the fame genus, and befides had them inclofedby double rows of hedges. The refult was, that eachof thefe plants produced great quantities of fertile feeds.Tournefort made the fame trial upon the Inpulus, Mil¬ler upon the bryony, and Geoffroy upon the may...	446>25>461>33>647 159>25>330>33>B O T A N Y <text>34>37>243>105>hemp, tranplanted them at a great distance from any males of the fame genus, and befides had them incolled by double rows of hedges. The reflust was, that each of thefe plants produced great quantities of fertile fehrs. Tournefort made the fame trial upon ...	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_01167	Normal	6	DAS (30 6Porn'll of Exeter, which fends two members to parlia¬ment: W. long. 40, and N. lat. 50° 25'.DARWENT, a river, which,' rifing in the Peak of Dar-byfhire, runs from north to fouth through that county,and falls into the Trent.DASYPUS, the Armadillo,in zoology, a genus ofqua¬drupeds belonging to the...	105>14>158>24>D A S 226>14>249>24>( 306 342>14>391>24>D A T <text>44>27>249>50>Fourth of Exeter, which fends two members to parlia› ment: W. long, 4°, and N. lat. 50° 25°. DARWENT, a river, which, riffing in the Peak of Dar› byhire, runs from north to fourth through that county,</text> <text>44>51>149>63>and falls into...	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_01556	Normal	733	C HI A m?1t mull'be acknowledged, That thefe two dlalles of wordsare io nearly allied to one another, that it is difficult toalcertain, in all cafes, the precife boundary betweenthem.Befides thefe, there are other words which fometimesaflame the province of pronouns, and- are generally con¬sidered as belonging to ...	105>15>127>22>G 130>15>151>22>TR 160>15>172>22>A 180>15>192>22>M 200>15>212>22>M 220>15>232>22>M 240>15>252>22>M 260>15>272>22>A 280>15>292>22>A 426>15>442>22>733 <text>14>25>225>54>It must be acknowledged, that thee two diflef of words are io nearly allied to one another, that it is difficult to are certain, in all ca...	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_01945	Normal	72	71 M E D Iamong the foldters, they are called camp-fevers ; in Hun¬gary, an Hungaric fever. But the pjague, or peftilence,is known when buboes and carbuncles arlfe in variousparts of the body. The fweating ficknefs had its rife inEngland, in which the patient fell into a violent fweat,of which many died in a day’s time...	55>25>65>31>72 171>22>234>31>M E D 253>22>340>31>C I N E <text>55>33>258>74>among the foldiers, they are called camp-fevers ; in Hurtin gary, an Hungaric fever. But the plague, or peltice, is known when bobs and carbunches arife in various parts of the body. The fweating ficknef has its rife in England, in which t...	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_02334	Normal		OPTgreat, apper.r"s from tl.e light of a candle; whrdi, ifthere be no obftacle in the way to obflrud the palTage ofits rays, will Gil all the fpace within two miles of thecfardle every way with luminous particles, before it haslofl: the lead: fenfible part of its fubdance.A ray of light is a continued dream of ihe...	157>8>232>17>O P T 423>8>442>17>S. <text>49>18>240>58>great, appears from the Light of a candle; which, if there be no obftacle in the way to obfture the pafagf of its rays, will fll all the fpace within two miles of the candle every way with luminous particles, before it has loft the leaf fenfible part of its fubance....	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]
brit_02723	Normal			0>0>500>500><other>	[{"model_id": "ds4sd/SmolDocling-256M-preview", "model_name": "SmolDocling-256M", "column_name": "markdown", "timestamp": "2026-06-05T14:15:53.878719", "output_format": "markdown", "max_tokens": 8192}]

Document Processing using SmolDocling-256M-preview

This dataset contains structured document extraction from images in davanstrien/ocr-affordances-pages using SmolDocling.

Processing Details

Source Dataset: davanstrien/ocr-affordances-pages
Model: ds4sd/SmolDocling-256M-preview
Number of Samples: 8
Processing Time: 1.9 minutes
Processing Date: 2026-06-05 14:15 UTC

Configuration

Image Column: image
Output Column: markdown
Output Format: markdown
Dataset Split: train
Batch Size: 32
Max Model Length: 8,192 tokens
Max Output Tokens: 8,192
GPU Memory Utilization: 80.0%

Model Information

SmolDocling-256M is an ultra-compact multimodal model that excels at:

💻 Code Recognition - Detects and formats code blocks with proper indentation
🔢 Formula Recognition - Identifies and processes mathematical expressions
📊 Tables & Charts - Extracts structured data from tables and charts
📐 Layout Preservation - Maintains document structure with bounding boxes
🏷️ DocTags Format - Efficient minimal representation for documents
⚡ Fast Inference - Only 256M parameters for quick processing

Dataset Structure

The dataset contains all original columns plus:

markdown: The extracted markdown from each image
inference_info: JSON list tracking all OCR models applied to this dataset

Usage

from datasets import load_dataset
import json



# Load the dataset
dataset = load_dataset("{output_dataset_id}", split="train")

# Access the extracted content
for example in dataset:
    
    print(example['markdown'])
    
    
    break

# View all OCR models applied to this dataset
inference_info = json.loads(dataset[0]["inference_info"])
for info in inference_info:
    print(f"Column: {info['column_name']} - Model: {info['model_id']}")

Reproduction

This dataset was generated using the uv-scripts/ocr SmolDocling script:

uv run https://huggingface.co/datasets/uv-scripts/ocr/raw/main/smoldocling-ocr.py \
    davanstrien/ocr-affordances-pages \
    <output-dataset> \
    --image-column image \
    --output-format markdown \
    --batch-size 32 \
    --max-model-len 8192 \
    --max-tokens 8192 \
    --gpu-memory-utilization 0.8

Performance

Processing Speed: ~0.1 images/second
Model Size: 256M parameters (ultra-compact)
GPU Configuration: vLLM with 80% GPU memory utilization

Generated with 🤖 UV Scripts

Downloads last month: 24

Number of rows:

Total file size:

5.71 MB