Skip to content
OpenMicroDrama

vaew /SkyScript-100M

Research dataset of 1B short-drama script and shooting-script pairs (sample only)
Datasets & papersMITEnglish README
Open on GitHubSave
Stars
150
30 days
0
Last push
1 year ago

About SkyScript-100M

SkyScript-100M pairs short-drama scripts with shooting scripts (scripts that describe scenes and camera shots). It comes from Huazhong University of Science and Technology and SkyWork AI. Only a small sample is public, so you can't download the full 1B pairs.

Best for: Dataset researchers who study short-drama script generation, shooting-script structure, or text-to-video training data.

What it does

  • Pairs 1,000,000,000 scripts with shooting scripts for short dramas
  • Built from ~6,660 short-drama titles, ~80,000 episodes, ~2,000 hours
  • Keyframe extraction and annotation produced ~10M shooting scripts
  • 100 script restorations per shooting script, made with the SkyReels model
  • Ships sample data in sample-en.json for early research
  • Technical report on arXiv (2408.09333v2)

Quickstart

Rewritten from the README. Check the repo for the latest steps.

  1. 1
    Open the README to read the dataset description.
  2. 2
    Download the sample file sample-en.json to see the data format.
  3. 3
    Read the technical report at https://arxiv.org/abs/2408.09333v2 for details.

Models and languages

Models it works with

  • SkyReels

Interface and docs

  • English

Alternatives

Other datasets & papers projects people compare with SkyScript-100M.

kangarooking

ai-short-drama-resources

Index of AI short-drama and comic-drama courses, in Chinese

Datasets & papers
51 GitHub stars2 months agoMIT

zhuyansen

awesome-claude-video-skills

Grades 204 agent skills that let coding agents make video

Datasets & papers
357 GitHub stars+10/30d2 days agoCC0-1.0

Vchitect

VBench

Scores video generation models on quality, faithfulness and safety

Datasets & papers
  • Gen-3
  • Kling
  • Pika
  • +1
1.8k GitHub stars+2/30d1 month agoApache-2.0

showlab

Awesome-Video-Diffusion

Video diffusion papers and projects, sorted into labeled sections

Datasets & papers
  • Wan 2.1
  • HunyuanVideo
  • LTX-Video
  • +7
5.8k GitHub stars+2/30d10 days agoNo license

Featured on OpenMicroDrama

Do you maintain SkyScript-100M? Add this badge to your README so English speakers can find our write-up.

Featured on OpenMicroDrama
markdown
[![Featured on OpenMicroDrama](https://openmicrodrama.com/badges/featured.svg)](https://openmicrodrama.com/projects/vaew-skyscript-100m)

The badge links to this page. Want your description changed? Email support@openmicrodrama.com.

Reviewed Oct 1, 2026. Last push 2024-11-16, outside our 90-day window. We kept it because it's still widely used. Only a small sample is released; the full 1B-pair corpus isn't openly downloadable. We write these descriptions ourselves. The repo's own docs are the final word.