Skip to content

Text-to-Video Generation

Score text-to-video output on two criteria

Visit official site

What Text-to-Video Generation is

The tool is described as a way to evaluate the quality of text-to-video generation, introducing a new evaluation metric called T2VScore.

What you can do with it

  • Evaluate the quality of text-to-video generation
  • Apply the T2VScore evaluation metric
  • Use the TVGE data set to assess and improve the metric
  • Judge videos on text-video alignment and video quality

The T2VScore metric

  • The metric combines two key criteria
  • Text-video alignment checks fidelity to the given text description
  • Video quality assesses the overall production level

TVGE data set

  • The data set collects human judgements on 2,543 text-to-video generated videos
  • Judgements cover the two criteria of the metric
  • The material states TVGE experiments show T2VScore provides a better evaluation metric

Pros & cons

✓ What we like

  • Combines alignment and quality in one metric
  • Provides a data set for evaluating and improving the metric

! What to watch out for

  • The metric is defined by two stated criteria
  • The TVGE judgements cover 2,543 generated videos

FAQ

What is T2VScore?

The material describes it as a new evaluation metric combining text-video alignment and video quality.

What is in the TVGE data set?

Human judgements on 2,543 text-to-video generated videos across the two criteria.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review