# Video Generation on Self-Forcing

Source: [Video Generation on Self-Forcing](<https://www.sota2.com/research/sota/video-generation-on-self-forcing>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-16T08:49:45.759Z

Evaluation of video generation models using the Self-Forcing dataset, assessed via perceptual metrics covering color consistency, subject fidelity, background coherence, text alignment, motion smoothness, and overall preference.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Video Generation |
| Dataset (including any stated split) | Self-Forcing |
| Metric | Color Consistency |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Color Consistency](<https://www.sota2.com/research/sota/video-generation-on-self-forcing.md?metric=Color%20Consistency>): Higher is better; unit: PERCENT.

- [Subject Consistency](<https://www.sota2.com/research/sota/video-generation-on-self-forcing.md?metric=Subject%20Consistency>): Higher is better; unit: PERCENT.

- [Background Consistency](<https://www.sota2.com/research/sota/video-generation-on-self-forcing.md?metric=Background%20Consistency>): Higher is better; unit: PERCENT.

- [Text Alignment](<https://www.sota2.com/research/sota/video-generation-on-self-forcing.md?metric=Text%20Alignment>): Higher is better; unit: PERCENT.

- [Motion Smoothness](<https://www.sota2.com/research/sota/video-generation-on-self-forcing.md?metric=Motion%20Smoothness>): Higher is better; unit: PERCENT.

- [Overall Preference](<https://www.sota2.com/research/sota/video-generation-on-self-forcing.md?metric=Overall%20Preference>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Color Consistency | Result date | Report paper |
| --- | --- | --- | --- |
| MemRoPE | 93 | Not recorded | 69b78e9d0636e5bf4afbdcc2 |
| MemRoPE | 90.4 | Not recorded | 69b78e9d0636e5bf4afbdcc2 |

### MemRoPE

Color Consistency: 93. Result ID: 69b7c41247f34ce9249f7eee.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e9d0636e5bf4afbdcc2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b78e9d0636e5bf4afbdcc2.md>)

Recorded evaluation configuration: {"Baseline":"Self-Forcing \[19\]"}. This record does not assert independent reproduction.

### MemRoPE

Color Consistency: 90.4. Result ID: 69b7c41247f34ce9249f7eef.

[Reporting paper](<https://www.sota2.com/research/paper/69b78e9d0636e5bf4afbdcc2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b78e9d0636e5bf4afbdcc2.md>)

Recorded evaluation configuration: {"Baseline":"Rolling Forcing \[27\]"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.