Skip to content

Dataset · FraunhoferIPK

Egocentric videos and manufacturing work scenarios

FraunhoferIPK/IndEgo

A multimodal dataset of first-person industrial scenarios and collaborative work for AI assistants.

What's inside

The repository contains data for visual question answering (VQA), summarization, and video classification tasks in English and German. The volume ranges from 10,000 to 100,000 examples. The dataset was created by researchers from the Fraunhofer IPK institute and a number of German universities to develop industrial assistants.

What you can do with it

  • Training multimodal AI assistants to support workers in manufacturing
  • Fine-tuning models for first-person video analysis and action question answering
Tasks
Visual question answering, Summarization, Video classification, Multimodal
Languages
English, German
Size
10K–100K rows
Downloads this month
61.4K
Likes
10

How to load it

from datasets import load_dataset

ds = load_dataset("FraunhoferIPK/IndEgo")