CoolFace
Datasetpublic

racineai/VDR_MEGA_2

VDR_MEGA_2 Dataset Summary VDR_MEGA_2 is a high-quality multimodal dataset created through the merge of multiple domain-specific datasets with enhanced data processing techniques. This dataset represents our most refined approach to multimodal data generation, incorporating filtering algorithms and improved AI-assisted content generation to deliver superior quality for RAG, DSE, question answering, document search, and vision-language model training tasks.… See the full description on the dataset page: https://huggingface.co/datasets/racineai/VDR_MEGA_2.

sourceHugging Faceapache-2.0updated 10mo agoView on Hugging Face
16likes8.9kdownloads
Dataset Card

VDRMEGA2

Dataset Summary

VDR_MEGA_2 is a high-quality multimodal dataset created through the merge of multiple domain-specific datasets with enhanced data processing techniques. This dataset represents our most refined approach to multimodal data generation, incorporating filtering algorithms and improved AI-assisted content generation to deliver superior quality for RAG, DSE, question answering, document search, and vision-language model training tasks.

Source Datasets

This merged dataset combines the the following datasets:

Dataset (split)DomainLanguage(s)
`racineai/VDR_Military (filtered)`MilitaryEN, FR
`racineai/VDR_Energy (filtered)`Energy/RegulationEN, FR
`racineai/VDR_Quantum (filtered)`QuantumEN, FR
`racineai/VDR_ibm-research_REAL-MM-RAG (train)`Technical/ResearchEN
`racineai/VDR_Cooking_Recipes (filtered)`Culinary ArtsMultiple
`racineai/VDR_Geotechnie (filtered)`GeotechnieEN, FR
`racineai/VDR_Nuclear (filtered)`Nuclear/RegulationEN, FR, DE, IT, ES
`racineai/VDR_2_vdr-visRAG-colpali (filtered)`VariousEN, FR, DE, IT, ES
`racineai/VDR_Renewable_Regulation (filtered)`Energy/RegulationMultiple
`racineai/VDR_Quantum_Circuit_Papers (filtered)`Quantum ComputingEN
`racineai/VDR_Hydrogen (filtered)`Hydrogen/RegulationEN, FR
`racineai/VDR_History_Geography (filtered)`Education/HistoryMultiple
`racineai/VDR_Energy_Arabic (train)`EnergyArabic
`racineai/VDR_CATIE-AQ_XMRec (train)`VariousFR

Data Fields

Each entry contains:

  • —`id` (string): Unique identifier
  • —`query` (string): High-quality technical/domain-specific question
  • —`image` (PIL.Image): High-resolution visual rendering of source document page
  • —`language` (string): Detected language of the image (queries sometimes differ on purpose)

Dataset Curators

  • —Léo Appourchaux
  • —Paul Lemaistre
  • —Yumeng Ye
  • —Mattéo KHAN
  • —André-Louis Rochet