README.md

metadata

language:
  - en
size_categories:
  - 10K<n<100K
task_categories:
  - summarization
  - image-to-text
  - text-generation
tags:
  - summarization
  - vision
  - DeepSeek-OCR
  - multilingual
  - visual-text-encoding
  - random-augmentation
library_name: datasets
license: cc-by-4.0
pretty_name: arXiv Scientific Paper Summarization
dataset_info:
  features:
    - name: text
      dtype: string
    - name: summary
      dtype: string
    - name: image
      dtype: image
    - name: source_dataset
      dtype: string
    - name: original_split
      dtype: string
    - name: original_index
      dtype: int64
  splits:
    - name: train
      num_examples: 49976

DeepSynth - arXiv Scientific Paper Summarization

Dataset Description

Scientific paper abstracts from arXiv, covering computer science, physics, and mathematics. Visual encoding preserves mathematical notation and document structure critical for scientific summarization.

This dataset is part of the DeepSynth project, which uses visual text encoding for multilingual summarization with the DeepSeek-OCR vision-language model. Text documents are converted into images and processed through a frozen 380M parameter visual encoder, enabling 20x token compression while preserving document layout and structure.

Key Features

• Original High-Quality Images: Full-resolution images stored once, augmented on-the-fly during training
• Random Augmentation Pipeline: Rotation, perspective, color jitter, and resize transforms for better generalization
• Visual Text Encoding: 20x compression ratio (1 visual token ≈ 20 text tokens)
• Document Structure Preservation: Layout and formatting maintained through image representation
• Human-Written Summaries: High-quality reference summaries for each document
• Deduplication Tracking: Source dataset and index tracking prevents duplicates

Dataset Statistics

• Total Samples: ~50,000
• Language(s): English
• Domain: Scientific papers (Computer Science, Physics, Mathematics)
• Average Document Length: ~5,000 tokens
• Average Summary Length: ~150 tokens

Source Dataset

Based on the arXiv dataset from arXiv.org e-prints.

• Papers: Computer Science, Physics, Mathematics domains
• Time Period: 2007-2021
• License: arXiv Non-exclusive distribution license

Image Augmentation Pipeline

Images are stored at original resolution (up to 1600×2200) and augmented during training for better generalization:

Available Augmentation Transforms

• Random Rotation: ±10° rotation for orientation invariance
• Random Perspective: 0.1-0.2 distortion to simulate viewing angles
• Random Resize: 512-1600px range for multi-scale learning
• Color Jitter: Brightness, contrast, saturation adjustments (±20%)
• Random Horizontal Flip: Optional (use with caution for text)

All transforms preserve aspect ratio with padding to maintain text readability. This approach:

• Reduces storage: 6x less disk space (single image vs 6 resolutions)
• Increases flexibility: Any resolution on-the-fly vs pre-computed fixed sizes
• Improves generalization: Random transforms prevent overfitting to specific resolutions

Dataset Structure

Data Fields

• text (string): Original document text
• summary (string): Human-written summary
• image (PIL.Image): Original full-size rendered document image (up to 1600×2200)
• source_dataset (string): Origin dataset name
• original_split (string): Source split (train/validation/test)
• original_index (int): Original sample index for deduplication

Data Example

{
    'text': 'We present a novel approach to neural machine translation...',
    'summary': 'This paper introduces an attention-based NMT model...',
    'image': <PIL.Image>,  # Original resolution (up to 1600×2200)
    'source_dataset': 'ccdv/arxiv-summarization',
    'original_split': 'train',
    'original_index': 0
}

Usage

Loading the Dataset

from datasets import load_dataset

# Load full dataset
dataset = load_dataset("baconnier/deepsynth-en-arxiv")

# Streaming for large datasets
dataset = load_dataset("baconnier/deepsynth-en-arxiv", streaming=True)

Training Example with DeepSeek-OCR and Augmentation

from transformers import AutoProcessor, AutoModelForVision2Seq
from datasets import load_dataset
from deepsynth.data.transforms import create_training_transform

# Load model and processor
model = AutoModelForVision2Seq.from_pretrained("deepseek-ai/DeepSeek-OCR")
processor = AutoProcessor.from_pretrained("deepseek-ai/DeepSeek-OCR")

# Load dataset
dataset = load_dataset("baconnier/deepsynth-en-arxiv")

# Create augmentation pipeline (random rotation, perspective, resize, color jitter)
transform = create_training_transform(
    target_size_range=(512, 1600),  # Random resize range
    rotation_degrees=10,             # ±10° rotation
    perspective_distortion=0.1,      # Perspective transform
    brightness_factor=0.2,           # ±20% brightness
    contrast_factor=0.2,             # ±20% contrast
)

# Process sample with augmentation
sample = dataset['train'][0]
augmented_image = transform(sample['image'])  # Apply random transforms
inputs = processor(
    images=augmented_image,
    text=sample['text'],
    return_tensors="pt"
)

# Fine-tune decoder only (freeze encoder)
for param in model.encoder.parameters():
    param.requires_grad = False

# Training loop with on-the-fly augmentation...

Training Recommendations

DeepSeek-OCR Fine-Tuning

# Recommended hyperparameters with augmentation
training_args = {
    "learning_rate": 2e-5,
    "batch_size": 4,
    "gradient_accumulation_steps": 4,
    "num_epochs": 3,
    "mixed_precision": "bf16",
    "freeze_encoder": True,  # IMPORTANT: Only fine-tune decoder

    # Augmentation parameters
    "rotation_degrees": 10,           # Random rotation ±10°
    "perspective_distortion": 0.1,    # Perspective transform
    "resize_range": (512, 1600),      # Random resize 512-1600px
    "brightness_factor": 0.2,         # ±20% brightness
    "contrast_factor": 0.2,           # ±20% contrast
}

Expected Performance

• Baseline (text-to-text): ROUGE-1 ~40-42
• DeepSeek-OCR (visual): ROUGE-1 ~44-47 (typical SOTA)
• Training Time: ~6-8 hours on A100 (80GB) for full dataset
• GPU Memory: ~40GB with batch_size=4, mixed_precision=bf16

Dataset Creation

This dataset was created using the DeepSynth pipeline:

1. Source Loading: Original text documents from ccdv/arxiv-summarization
2. Text-to-Image Conversion: Documents rendered as PNG images (DejaVu Sans 12pt, Unicode support)
3. Original Resolution Storage: Full-quality images stored once (up to 1600×2200)
4. Incremental Upload: Batches of 5,000 samples uploaded to HuggingFace Hub
5. Deduplication: Source tracking prevents duplicate samples

Note: Images are augmented on-the-fly during training using random transformations (rotation, perspective, resize, color jitter) for better generalization across different resolutions and conditions.

Rendering Specifications

• Font: DejaVu Sans 12pt (full Unicode support for multilingual text)
• Line Wrapping: 100 characters per line
• Margin: 40px
• Background: White (255, 255, 255)
• Text Color: Black (0, 0, 0)
• Format: PNG with lossless compression

Citation

If you use this dataset in your research, please cite:

@misc{deepsynth-en-arxiv,
    title={{DeepSynth arXiv Scientific Paper Summarization: Visual Text Encoding with Random Augmentation for Summarization}},
    author={Baconnier},
    year={2025},
    publisher={HuggingFace},
    howpublished={\url{https://huggingface.co/datasets/baconnier/deepsynth-en-arxiv}}
}

Source Dataset Citation

@article{cohan2018discourse,
    title={A discourse-aware attention model for abstractive summarization of long documents},
    author={Cohan, Arman and Dernoncourt, Franck and Kim, Doo Soon and Bui, Trung and Kim, Seokhwan and Chang, Walter and Goharian, Nazli},
    journal={arXiv preprint arXiv:1804.05685},
    year={2018}
}

License

Creative Commons Attribution 4.0 International (CC BY 4.0)

Note: This dataset inherits the license from the original source dataset. Please review the source license before commercial use.

Limitations and Bias

• Scientific jargon: Heavy use of technical terminology
• Mathematical notation: LaTeX rendering may affect OCR accuracy
• Domain-specific: Optimized for CS/Physics/Math papers
• Length: Very long documents (up to 10,000+ tokens) may be truncated

Additional Information

Dataset Curators

Created by the DeepSynth team as part of multilingual visual summarization research.

Contact

• Repository: DeepSynth GitHub
• Issues: GitHub Issues

Acknowledgments

• DeepSeek-OCR: Visual encoder from DeepSeek AI
• Source Dataset: ccdv/arxiv-summarization
• HuggingFace: Dataset hosting and infrastructure