Skip to content

Dataset · OpenGVLab

10M video-caption pairs InternVid-10M-FLT

OpenGVLab/InternVid

10 million video clips with automatically generated captions and UMT scores for video-text tasks.

What's inside

A subset of InternVid-10M-FLT: 10 million video clips from public web videos with generated high-quality English captions. Data is stored in JSONL format; columns include videoID, timestamps, caption, and UMT similarity score. The card references loading via load_dataset.

What you can do with it

  • Training video-text models (e.g., CLIP-like models)
  • Extracting video features
  • Generating video captions
Tasks
Embeddings
Languages
English
Size
10M–100M rows
Downloads this month
438
Likes
101
Access
on request on Hugging Face

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("OpenGVLab/InternVid")