Skip to content

Dataset · omegalabsinc

Multimodal video dataset from the Bittensor network

omegalabsinc/omega-multimodal

A massive video corpus with text descriptions and tabular metadata from the decentralized OMEGA Labs network.

What's inside

The dataset was collected by the decentralized Bittensor network and combines video recordings, texts, and tabular data. The authors claim over 1 million hours of video footage and more than 30 million two-minute clips. The collection covers a wide range of user-generated content for video analytics and multimodal learning tasks.

What you can do with it

  • Training video-to-text generation models
  • Training video and image classifiers
  • Creating image-to-video generation systems and extracting visual features
Tasks
Video text → text, Video classification, Image classification, Image → text, Image → video, Image feature extraction
Downloads this month
11.3K
Likes
61

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("omegalabsinc/omega-multimodal")