# Text-to-3D Avatar Generation on 50 Text Prompts

Source: [Text-to-3D Avatar Generation on 50 Text Prompts](<https://www.sota2.com/research/sota/text-to-3d-avatar-generation-on-50-text-prompts>)

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

Data updated: 2026-04-10T01:01:54.899Z

Evaluation of Text-to-3D Avatar Generation models using a specific set of 50 text prompts, assessed using perceptual metrics including BLIP-VQA, Text CLIP Score, HyperIQA, and FID.

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 | Text-to-3D Avatar Generation |
| Dataset (including any stated split) | 50 Text Prompts |
| Metric | BLIP-VQA |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [BLIP-VQA](<https://www.sota2.com/research/sota/text-to-3d-avatar-generation-on-50-text-prompts.md?metric=BLIP-VQA>): Higher is better; unit: SCALAR.

- [Text CLIP Score](<https://www.sota2.com/research/sota/text-to-3d-avatar-generation-on-50-text-prompts.md?metric=Text%20CLIP%20Score>): Higher is better; unit: SCALAR.

- [CLIB-FIQA](<https://www.sota2.com/research/sota/text-to-3d-avatar-generation-on-50-text-prompts.md?metric=CLIB-FIQA>): Higher is better; unit: SCALAR.

- [HyperIQA](<https://www.sota2.com/research/sota/text-to-3d-avatar-generation-on-50-text-prompts.md?metric=HyperIQA>): Higher is better; unit: PERCENT.

- [FID (2D Diffusion)](<https://www.sota2.com/research/sota/text-to-3d-avatar-generation-on-50-text-prompts.md?metric=FID%20(2D%20Diffusion)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | BLIP-VQA | Result date | Report paper |
| --- | --- | --- | --- |
| GenLCA | 0.64 | Not recorded | 69d7329d7d0172140ec76bf9 |
| HumanGaussian | 0.62 | Not recorded | 69d7329d7d0172140ec76bf9 |
| DreamWaltz-G | 0.58 | Not recorded | 69d7329d7d0172140ec76bf9 |
| TADA | 0.5 | Not recorded | 69d7329d7d0172140ec76bf9 |
| TeRA | 0.42 | Not recorded | 69d7329d7d0172140ec76bf9 |
| SIGMAN | 0.29 | Not recorded | 69d7329d7d0172140ec76bf9 |

### GenLCA

BLIP-VQA: 0.64. Result ID: 69d84be847f34ce924c09e4f.

[Reporting paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

Recorded evaluation configuration: {"Inference Time":"12s"}. This record does not assert independent reproduction.

### HumanGaussian

BLIP-VQA: 0.62. Result ID: 69d84be847f34ce924c09e4b.

[Reporting paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72851e209131cea56fb4.md>)

[Implementation](<https://alvinliu0.github.io/projects/HumanGaussian>)

Recorded evaluation configuration: {"Inference Time":"1.2h"}. This record does not assert independent reproduction.

### DreamWaltz-G

BLIP-VQA: 0.58. Result ID: 69d84be847f34ce924c09e4c.

[Reporting paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74fa1e209131cea650db.md>)

Recorded evaluation configuration: {"Inference Time":"3.0h"}. This record does not assert independent reproduction.

### TADA

BLIP-VQA: 0.5. Result ID: 69d84be847f34ce924c09e4a.

[Reporting paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70621e209131cea4ac90.md>)

Recorded evaluation configuration: {"Inference Time":"2.5h"}. This record does not assert independent reproduction.

### TeRA

BLIP-VQA: 0.42. Result ID: 69d84be847f34ce924c09e4d.

[Reporting paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b201e209131cea86b39.md>)

Recorded evaluation configuration: {"Inference Time":"12s"}. This record does not assert independent reproduction.

### SIGMAN

BLIP-VQA: 0.29. Result ID: 69d84be847f34ce924c09e4e.

[Reporting paper](<https://www.sota2.com/research/paper/69d7329d7d0172140ec76bf9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77f01e209131cea7623e.md>)

Recorded evaluation configuration: {"Inference Time":"3s"}. This record does not assert independent reproduction.

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