Skip to content

Dataset · Rapidata

Sora video visual style Likert scores

Rapidata/sora-video-generation-style-likert-scoring

198 generated videos with 6,021 human Likert scale ratings for visual appeal.

What's inside

This English-language multimodal dataset (tables, text, video) contains 198 clips and 6,021 assessor votes collected in about an hour via the Rapidata API. Parquet files record human Likert scale ratings evaluating the visual style of the generated videos.

What you can do with it

  • Training reward models to evaluate the visual appeal of generated videos
  • Validating and comparing text-to-video generation models based on human preferences
Tasks
Video classification, Text → video
Languages
English
Size
up to 1K rows
Downloads this month
382
Likes
34

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("Rapidata/sora-video-generation-style-likert-scoring")