# Co-speech gesture generation on BEAT2 All 25 Speakers

Source: [Co-speech gesture generation on BEAT2 All 25 Speakers](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat2-all-25-speakers>)

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

Data updated: 2026-05-27T07:50:23.292Z

Evaluation of Co-speech gesture generation models on the full BEAT2 dataset (all 25 speakers), utilizing quantitative metrics including FGD, BA, Diversity, MSE, and LVD.

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 | Co-speech gesture generation |
| Dataset (including any stated split) | BEAT2 All 25 Speakers |
| Metric | FGD |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [FGD](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat2-all-25-speakers.md?metric=FGD>): Higher is better; unit: SCALAR.

- [BA](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat2-all-25-speakers.md?metric=BA>): Higher is better; unit: SCALAR.

- [Diversity](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat2-all-25-speakers.md?metric=Diversity>): Higher is better; unit: SCALAR.

- [MSE](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat2-all-25-speakers.md?metric=MSE>): Lower is better; unit: SCALAR.

- [LVD](<https://www.sota2.com/research/sota/co-speech-gesture-generation-on-beat2-all-25-speakers.md?metric=LVD>): Lower is better; unit: SCALAR.

## Recorded results

| Method | FGD | Result date | Report paper |
| --- | --- | --- | --- |
| TalkSHOW | 6.145 | Not recorded | 6a167aa0dbd7150506f7fff8 |
| EMAGE | 5.643 | Not recorded | 6a167aa0dbd7150506f7fff8 |
| SemTalk | 5.214 | Not recorded | 6a167aa0dbd7150506f7fff8 |
| GestureLSM | 4.268 | Not recorded | 6a167aa0dbd7150506f7fff8 |
| DuoGesture | 4.081 | Not recorded | 6a167aa0dbd7150506f7fff8 |

### TalkSHOW

FGD: 6.145. Result ID: 6a16a1a3fbc373592c0f8380.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa0dbd7150506f7fff8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fcb1e209131cea4771a.md>)

[Implementation](<https://talkshow.is.tue.mpg.de/>)

Recorded evaluation configuration: {"Setting":"All Speakers"}. This record does not assert independent reproduction.

### EMAGE

FGD: 5.643. Result ID: 6a16a1a3fbc373592c0f8382.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa0dbd7150506f7fff8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72c11e209131cea5867d.md>)

[Implementation](<https://pantomatrix.github.io/EMAGE/>)

Recorded evaluation configuration: {"Setting":"All Speakers"}. This record does not assert independent reproduction.

### SemTalk

FGD: 5.214. Result ID: 6a16a1a3fbc373592c0f8383.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa0dbd7150506f7fff8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77591e209131cea7292d.md>)

[Implementation](<https://xiangyue-zhang.github.io/SemTalk>)

Recorded evaluation configuration: {"Setting":"All Speakers"}. This record does not assert independent reproduction.

### GestureLSM

FGD: 4.268. Result ID: 6a16a1a3fbc373592c0f8381.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa0dbd7150506f7fff8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a641e209131cea82d92.md>)

[Implementation](<https://andypinxinliu.github.io/GestureLSM>)

Recorded evaluation configuration: {"Setting":"All Speakers"}. This record does not assert independent reproduction.

### DuoGesture

FGD: 4.081. Result ID: 6a16a1a3fbc373592c0f8384.

[Reporting paper](<https://www.sota2.com/research/paper/6a167aa0dbd7150506f7fff8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a167aa0dbd7150506f7fff8.md>)

Recorded evaluation configuration: {"Setting":"All Speakers"}. This record does not assert independent reproduction.

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