HyperAIHyperAI

Command Palette

Search for a command to run...

Console

WIT image-text Dataset

Date

3 years ago

Size

25.2 GB

Organization

Google

Publish URL

github.com

Paper URL

arxiv.org

License

Other

Tags

Multimodality
Natural Language Processing
Dataset Download
Join the Discord Community
Featured Image

WIT, short for Wikipedia-based Image Text, is a large multimodal and multilingual dataset. The dataset consists of a curated collection of 37.6 million entity-rich image-text examples, including 11.5 million unique images in 108 Wikipedia languages. The scale of the dataset allows it to be used as a pre-training dataset for multimodal machine learning models.

WIT has four unique advantages:

  • WIT is the largest multimodal dataset in terms of the number of image-text examples.
  • Over 100 languages are covered (with at least 12,000 examples per language), and cross-lingual text is provided for many images.
  • Relative to previous datasets, WIT represents a more diverse set of concepts and real-world entities.
  • WIT provides a very challenging real-world test set.
WIT.torrent
Seeding 1Downloading 1Completed 574Total Downloads 741
  • WIT/
    • README.md
      1.41 KB
    • README.txt
      2.82 KB
      • data/
          • test/
            • wit_v1.test.all-00000-of-00005.tsv.gz
              30.8 MB
            • wit_v1.test.all-00001-of-00005.tsv.gz
              61.46 MB
            • wit_v1.test.all-00002-of-00005.tsv.gz
              91.95 MB
            • wit_v1.test.all-00003-of-00005.tsv.gz
              122.5 MB
            • wit_v1.test.all-00004-of-00005.tsv.gz
              153.14 MB
          • train/
            • wit_v1.train.all-00000-of-00010.tsv.gz
              2.64 GB
            • wit_v1.train.all-00001-of-00010.tsv.gz
              5.12 GB
            • wit_v1.train.all-00002-of-00010.tsv.gz
              7.61 GB
            • wit_v1.train.all-00003-of-00010.tsv.gz
              10.1 GB
            • wit_v1.train.all-00004-of-00010.tsv.gz
              12.58 GB
            • wit_v1.train.all-00005-of-00010.tsv.gz
              15.07 GB
            • wit_v1.train.all-00006-of-00010.tsv.gz
              17.56 GB
            • wit_v1.train.all-00007-of-00010.tsv.gz
              20.04 GB
            • wit_v1.train.all-00008-of-00010.tsv.gz
              22.53 GB
            • wit_v1.train.all-00009-of-00010.tsv.gz
              25.02 GB
          • val/
            • wit_v1.val.all-00000-of-00005.tsv.gz
              25.05 GB
            • wit_v1.val.all-00001-of-00005.tsv.gz
              25.09 GB
            • wit_v1.val.all-00002-of-00005.tsv.gz
              25.13 GB
            • wit_v1.val.all-00003-of-00005.tsv.gz
              25.17 GB
            • wit_v1.val.all-00004-of-00005.tsv.gz
              25.2 GB

Build AI with AI

From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.

AI Co-coding
Ready-to-use GPUs
Best Pricing
Get Started

Hyper Newsletters

Subscribe to our latest updates
We will deliver the latest updates of the week to your inbox at nine o'clock every Monday morning
Powered by MailChimp
WIT image-text Dataset | Datasets | HyperAI