Datasets:

davanstrien
/

ocr-smoke-paddleocr-v1-stable

raw_queries listlengths 1 1	broad_topical_query stringclasses 3 values	broad_topical_explanation stringclasses 3 values	specific_detail_query stringclasses 3 values	specific_detail_explanation stringclasses 3 values	visual_element_query stringclasses 3 values	visual_element_explanation stringclasses 3 values	parsed_into_json bool 1 class	markdown stringclasses 3 values	inference_info stringclasses 1 value
[ "{\n \"broad_topical_query\": \"UFO FlyBys in Middle Tennessee\",\n \"broad_topical_explanation\": \"This query targets the main subject of the document, which is the reported UFO sightings in Middle Tennessee.\",\n \"specific_detail_query\": \"David Copperfield special\",\n \"specific_detail_explanation\": \"T...	UFO FlyBys in Middle Tennessee	This query targets the main subject of the document, which is the reported UFO sightings in Middle Tennessee.	David Copperfield special	This query focuses on a specific detail mentioned in the document, which is the David Copperfield special that the author and her husband were watching.	Image of a man and woman	This query references the visual element in the document, which is the image of a man and woman sitting on the porch.	true	MUSTALL SONG, Muthilla, TV - At 2, N 1995 Q: 3. United OUT OF THIS WORLD UFO FlyBys in Middle Tennessee BY JOYSA M. WINTER aVere Piaut is well aware that some people out there think she is least aware people think she's nuts. Her story is so bizarre, in fact, that sometimes Piaut herself can barely believe it. It was a...	[{"model_id": "PaddlePaddle/PaddleOCR-VL", "model_name": "PaddleOCR-VL", "model_size": "0.9B", "task_mode": "ocr", "column_name": "markdown", "timestamp": "2026-02-17T10:48:33.815988", "temperature": 0.0, "max_tokens": 4096, "smart_resize": true}]
[ "{\n \"broad_topical_query\": \"Snake with legs and feet reported in Africa\",\n \"broad_topical_explanation\": \"This query focuses on the main subject of the document, which is the report of a snake with legs and feet found in Africa.\",\n \"specific_detail_query\": \"Snake with legs and feet reported in Afric...	Snake with legs and feet reported in Africa	This query focuses on the main subject of the document, which is the report of a snake with legs and feet found in Africa.	Snake with legs and feet reported in Africa, 1899	This query is more specific, focusing on the date and location mentioned in the document.	Document page with text	This query refers to the visual element of the document page, which includes the text content and layout.	true	A Waco Snake. Waco Telephone: J. W. Boynton brought to the Telephone yesterday afternoon was his first call with a teammate and the existence of which has been denied by some people. It was nothing more nor less than a snake with legs of small size which was small one, not more than sixteen inches long. was what is kno...	[{"model_id": "PaddlePaddle/PaddleOCR-VL", "model_name": "PaddleOCR-VL", "model_size": "0.9B", "task_mode": "ocr", "column_name": "markdown", "timestamp": "2026-02-17T10:48:33.815988", "temperature": 0.0, "max_tokens": 4096, "smart_resize": true}]
[ "{\n \"broad_topical_query\": \"Document about Mount Clemens, Michigan\",\n \"broad_topical_explanation\": \"This query is effective because it covers the main subject of the document, which is the investigation of a UFO sighting in Mount Clemens, Michigan.\",\n \"specific_detail_query\": \"Document mentioning D...	Document about Mount Clemens, Michigan	This query is effective because it covers the main subject of the document, which is the investigation of a UFO sighting in Mount Clemens, Michigan.	Document mentioning Dr. Hynek and the OVNI	This query is effective because it focuses on a specific detail from the document, which is the involvement of Dr. Hynek in the investigation of the OVNI.	Document with a picture of Paris 67a	This query is effective because it references a visual element in the document, which is the picture of Paris 67a, and can help in retrieving the document.	true	M o u n t Clemens, Michigan, cerca delago St. George, C.B. Base. Aé- reade Self-Comoas y os- seraron un objeto el lago del sol, de sobre el lago del sol, de 10 minutos. En ese momento, el 14 fotos con una cámara Polaroid Swinger, aé- reade, y el cuyato que se duro, duro mucho mayor" que la de cualquier aporte a convenc...	[{"model_id": "PaddlePaddle/PaddleOCR-VL", "model_name": "PaddleOCR-VL", "model_size": "0.9B", "task_mode": "ocr", "column_name": "markdown", "timestamp": "2026-02-17T10:48:33.815988", "temperature": 0.0, "max_tokens": 4096, "smart_resize": true}]

Document Processing using PaddleOCR-VL (OCR mode)

This dataset contains OCR results from images in davanstrien/ufo-ColPali using PaddleOCR-VL, an ultra-compact 0.9B OCR model.

Processing Details

Source Dataset: davanstrien/ufo-ColPali
Model: PaddlePaddle/PaddleOCR-VL
Task Mode: ocr - General text extraction to markdown format
Number of Samples: 3
Processing Time: 1.2 min
Processing Date: 2026-02-17 10:48 UTC

Configuration

Image Column: image
Output Column: markdown
Dataset Split: train
Batch Size: 16
Smart Resize: Enabled
Max Model Length: 8,192 tokens
Max Output Tokens: 4,096
Temperature: 0.0
GPU Memory Utilization: 80.0%

Model Information

PaddleOCR-VL is a state-of-the-art, resource-efficient model tailored for document parsing:

🎯 Ultra-compact - Only 0.9B parameters (smallest OCR model)
📝 OCR mode - General text extraction
📊 Table mode - HTML table recognition
📐 Formula mode - LaTeX mathematical notation
📈 Chart mode - Structured chart analysis
🌍 Multilingual - Support for multiple languages
⚡ Fast - Quick initialization and inference
🔧 ERNIE-4.5 based - Different architecture from Qwen models

Task Modes

OCR: Extract text content to markdown format
Table Recognition: Extract tables to HTML format
Formula Recognition: Extract mathematical formulas to LaTeX
Chart Recognition: Analyze and describe charts/diagrams

Dataset Structure

The dataset contains all original columns plus:

markdown: The extracted content based on task mode
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"Task: {info['task_mode']} - Model: {info['model_id']}")

Reproduction

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

uv run https://huggingface.co/datasets/uv-scripts/ocr/raw/main/paddleocr-vl.py \
    davanstrien/ufo-ColPali \
    <output-dataset> \
    --task-mode ocr \
    --image-column image \
    --batch-size 16 \
    --max-model-len 8192 \
    --max-tokens 4096 \
    --gpu-memory-utilization 0.8

Performance

Model Size: 0.9B parameters (smallest among OCR models)
Processing Speed: ~0.04 images/second
Architecture: NaViT visual encoder + ERNIE-4.5-0.3B language model

Generated with 🤖 UV Scripts

Downloads last month: 15

Number of rows:

Total file size:

328 kB