# Text-to-Image Synthesis on Human evaluation dataset 500 texts

Source: [Text-to-Image Synthesis on Human evaluation dataset 500 texts](<https://www.sota2.com/research/sota/text-to-image-synthesis-on-human-evaluation-dataset-500-texts>)

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

Data updated: 2026-02-26T05:46:59.735Z

Human evaluation of Text-to-Image Synthesis models based on 500 samples, assessed using criteria like Image Clarity, Texture Quality, and Relevance to the Text.

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-Image Synthesis |
| Dataset (including any stated split) | Human evaluation dataset 500 texts |
| Metric | Image Clarity |
| Unit | SCALAR |
| 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

- [Image Clarity](<https://www.sota2.com/research/sota/text-to-image-synthesis-on-human-evaluation-dataset-500-texts.md?metric=Image%20Clarity>): Higher is better; unit: SCALAR.

- [Texture Quality](<https://www.sota2.com/research/sota/text-to-image-synthesis-on-human-evaluation-dataset-500-texts.md?metric=Texture%20Quality>): Higher is better; unit: SCALAR.

- [Text Relevance](<https://www.sota2.com/research/sota/text-to-image-synthesis-on-human-evaluation-dataset-500-texts.md?metric=Text%20Relevance>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Image Clarity | Result date | Report paper |
| --- | --- | --- | --- |
| ERNIE-ViLG | 4.221 | Not recorded | 68fa6b871e209131cea2eddf |
| CogView | 3.867 | Not recorded | 68fa6b871e209131cea2eddf |

### ERNIE-ViLG

Image Clarity: 4.221. Result ID: 698e947dffc3901de09ed45e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b871e209131cea2eddf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b871e209131cea2eddf.md>)

Recorded evaluation configuration: {"Setting":"zero-shot"}. This record does not assert independent reproduction.

### CogView

Image Clarity: 3.867. Result ID: 698e947dffc3901de09ed45d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b871e209131cea2eddf.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b2b1e209131cea2c82b.md>)

[Implementation](<https://github.com/THUDM/CogView>)

Recorded evaluation configuration: {"Setting":"zero-shot"}. This record does not assert independent reproduction.

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