# Text-to-Image Generation on Parti-Prompts (test)

Source: [Text-to-Image Generation on Parti-Prompts (test)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test>)

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

Data updated: 2026-05-07T08:56:09.762Z

Evaluation of Text-to-Image generation models using the PartiPrompts test set, assessed using automated quality metrics including Aesthetic Score, PickScore, HPS, and CLIP scores (reported as Median/Mean).

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 Generation |
| Dataset (including any stated split) | Parti-Prompts (test) |
| Metric | Aesthetic Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 33 of 33 |
| 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

- [Aesthetic Score](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=Aesthetic%20Score>): Higher is better; unit: SCALAR.

- [Aesthetic Score (Median)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=Aesthetic%20Score%20(Median)>): Higher is better; unit: SCALAR.

- [Aesthetic Score (Mean)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=Aesthetic%20Score%20(Mean)>): Higher is better; unit: SCALAR.

- [PickScore (Median)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=PickScore%20(Median)>): Higher is better; unit: SCALAR.

- [PickScore (Mean)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=PickScore%20(Mean)>): Higher is better; unit: SCALAR.

- [HPS (Median)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=HPS%20(Median)>): Higher is better; unit: SCALAR.

- [HPS (Mean)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=HPS%20(Mean)>): Higher is better; unit: SCALAR.

- [CLIP Score (Median)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=CLIP%20Score%20(Median)>): Higher is better; unit: SCALAR.

- [CLIP Score (Mean)](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=CLIP%20Score%20(Mean)>): Higher is better; unit: SCALAR.

- [PickScore](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=PickScore>): Higher is better; unit: SCALAR.

- [HPS](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=HPS>): Higher is better; unit: SCALAR.

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

- [HPS v2](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=HPS%20v2>): Higher is better; unit: SCALAR.

- [HPS v3](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=HPS%20v3>): Higher is better; unit: SCALAR.

- [IR Score](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=IR%20Score>): Higher is better; unit: SCALAR.

- [Average Score](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=Average%20Score>): Higher is better; unit: SCALAR.

- [HPS v2.1](<https://www.sota2.com/research/sota/text-to-image-generation-on-parti-prompts-test.md?metric=HPS%20v2.1>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Aesthetic Score | Result date | Report paper |
| --- | --- | --- | --- |
| PGD | 88.4 | Not recorded | 699d30636017c4460ffab21e |
| SPO-SDXL | 88.1 | Not recorded | 699d30636017c4460ffab21e |
| cPGD | 83.6 | Not recorded | 699d30636017c4460ffab21e |
| PGD | 77.8 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SD1.5 | 74.63 | Not recorded | 68fa77911e209131cea73f9f |
| MaPO-SDXL | 72.4 | Not recorded | 699d30636017c4460ffab21e |
| cPGD | 70.1 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SDXL | 69.7 | Not recorded | 68fa77911e209131cea73f9f |
| PGD | 68.8 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SD1.5 | 68.01 | Not recorded | 68fa77911e209131cea73f9f |
| PGD | 67.6 | Not recorded | 699d30636017c4460ffab21e |
| NPO | 63.4 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SDXL | 61.7 | Not recorded | 68fa77911e209131cea73f9f |
| cPGD | 59.4 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SD1.5 | 59.01 | Not recorded | 68fa77911e209131cea73f9f |
| DPO-SDXL | 57.5 | Not recorded | 699d30636017c4460ffab21e |
| cPGD | 57.2 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SDXL | 57.17 | Not recorded | 68fa77911e209131cea73f9f |
| NPO | 55.1 | Not recorded | 699d30636017c4460ffab21e |
| InPO-SD1.5 | 52.27 | Not recorded | 68fa77911e209131cea73f9f |
| SDXL | 50 | Not recorded | 699d30636017c4460ffab21e |
| TGO | 5.51 | Not recorded | 69fc1cb45497b3b670c886b3 |
| SFT | 5.48 | Not recorded | 69fc1cb45497b3b670c886b3 |
| Diffusion-KTO | 5.47 | Not recorded | 69fc1cb45497b3b670c886b3 |
| CSFT | 5.45 | Not recorded | 69fc1cb45497b3b670c886b3 |
| Diffusion-DPO | 5.32 | Not recorded | 69fc1cb45497b3b670c886b3 |
| SD v1.5 | 5.25 | Not recorded | 69fc1cb45497b3b670c886b3 |
| AlignProp | 5.25 | Not recorded | 69fc1cb45497b3b670c886b3 |
| DSPO | 5.25 | Not recorded | 69fc1cb45497b3b670c886b3 |

### PGD

Aesthetic Score: 88.4. Result ID: 699e2b2a5698665ec4e6318d.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"SPO-SDXL","Inference Strategy":"PGD"}. This record does not assert independent reproduction.

### SPO-SDXL

Aesthetic Score: 88.1. Result ID: 699e2b2a5698665ec4e6318c.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77921e209131cea74088.md>)

Recorded evaluation configuration: {"Base Model":"SPO-SDXL","Inference Strategy":"-"}. This record does not assert independent reproduction.

### cPGD

Aesthetic Score: 83.6. Result ID: 699e2b2a5698665ec4e6318e.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"SPO-SDXL","Inference Strategy":"cPGD"}. This record does not assert independent reproduction.

### PGD

Aesthetic Score: 77.8. Result ID: 699e2b2a5698665ec4e6318a.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"MaPO-SDXL","Inference Strategy":"PGD"}. This record does not assert independent reproduction.

### InPO-SD1.5

Aesthetic Score: 74.63. Result ID: 699540d93badfaffe62b85d6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"Base-SD1.5"}. This record does not assert independent reproduction.

### MaPO-SDXL

Aesthetic Score: 72.4. Result ID: 699e2b2a5698665ec4e63189.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73891e209131cea5d0f2.md>)

Recorded evaluation configuration: {"Base Model":"MaPO-SDXL","Inference Strategy":"-"}. This record does not assert independent reproduction.

### cPGD

Aesthetic Score: 70.1. Result ID: 699e2b2a5698665ec4e6318b.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"MaPO-SDXL","Inference Strategy":"cPGD"}. This record does not assert independent reproduction.

### InPO-SDXL

Aesthetic Score: 69.7. Result ID: 699540d93badfaffe62b85d4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"SFT-SDXL"}. This record does not assert independent reproduction.

### PGD

Aesthetic Score: 68.8. Result ID: 699e2b2a5698665ec4e63183.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"SDXL","Inference Strategy":"PGD"}. This record does not assert independent reproduction.

### InPO-SD1.5

Aesthetic Score: 68.01. Result ID: 699540d93badfaffe62b85d8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"DPO-SD1.5"}. This record does not assert independent reproduction.

### PGD

Aesthetic Score: 67.6. Result ID: 699e2b2a5698665ec4e63187.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"DPO-SDXL","Inference Strategy":"PGD"}. This record does not assert independent reproduction.

### NPO

Aesthetic Score: 63.4. Result ID: 699e2b2a5698665ec4e63186.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa789f1e209131cea7968b.md>)

Recorded evaluation configuration: {"Base Model":"DPO-SDXL","Inference Strategy":"NPO"}. This record does not assert independent reproduction.

### InPO-SDXL

Aesthetic Score: 61.7. Result ID: 699540d93badfaffe62b85d3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"Base-SDXL"}. This record does not assert independent reproduction.

### cPGD

Aesthetic Score: 59.4. Result ID: 699e2b2a5698665ec4e63188.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"DPO-SDXL","Inference Strategy":"cPGD"}. This record does not assert independent reproduction.

### InPO-SD1.5

Aesthetic Score: 59.01. Result ID: 699540d93badfaffe62b85d9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"KTO-SD1.5"}. This record does not assert independent reproduction.

### DPO-SDXL

Aesthetic Score: 57.5. Result ID: 699e2b2a5698665ec4e63185.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa713d1e209131cea4ffe9.md>)

Recorded evaluation configuration: {"Base Model":"DPO-SDXL","Inference Strategy":"-"}. This record does not assert independent reproduction.

### cPGD

Aesthetic Score: 57.2. Result ID: 699e2b2a5698665ec4e63184.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

Recorded evaluation configuration: {"Base Model":"SDXL","Inference Strategy":"cPGD"}. This record does not assert independent reproduction.

### InPO-SDXL

Aesthetic Score: 57.17. Result ID: 699540d93badfaffe62b85d5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"DPO-SDXL"}. This record does not assert independent reproduction.

### NPO

Aesthetic Score: 55.1. Result ID: 699e2b2a5698665ec4e63182.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa789f1e209131cea7968b.md>)

Recorded evaluation configuration: {"Base Model":"SDXL","Inference Strategy":"NPO"}. This record does not assert independent reproduction.

### InPO-SD1.5

Aesthetic Score: 52.27. Result ID: 699540d93badfaffe62b85d7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77911e209131cea73f9f.md>)

[Implementation](<https://github.com/JaydenLyh/InPO>)

Recorded evaluation configuration: {"baseline\_comparison":"SFT-SD1.5"}. This record does not assert independent reproduction.

### SDXL

Aesthetic Score: 50. Result ID: 699e2b2a5698665ec4e63181.

[Reporting paper](<https://www.sota2.com/research/paper/699d30636017c4460ffab21e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70011e209131cea487fb.md>)

Recorded evaluation configuration: {"Base Model":"SDXL","Inference Strategy":"-"}. This record does not assert independent reproduction.

### TGO

Aesthetic Score: 5.51. Result ID: 69fc539a47f34ce924df27dd.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### SFT

Aesthetic Score: 5.48. Result ID: 69fc539a47f34ce924df27d7.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### Diffusion-KTO

Aesthetic Score: 5.47. Result ID: 69fc539a47f34ce924df27db.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75371e209131cea6682e.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### CSFT

Aesthetic Score: 5.45. Result ID: 69fc539a47f34ce924df27d8.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70b41e209131cea4caeb.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### Diffusion-DPO

Aesthetic Score: 5.32. Result ID: 69fc539a47f34ce924df27da.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa713d1e209131cea4ffe9.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### SD v1.5

Aesthetic Score: 5.25. Result ID: 69fc539a47f34ce924df27d6.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c391e209131cea32e4a.md>)

[Implementation](<https://github.com/CompVis/latent-diffusion>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### AlignProp

Aesthetic Score: 5.25. Result ID: 69fc539a47f34ce924df27d9.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa759b1e209131cea68eed.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

### DSPO

Aesthetic Score: 5.25. Result ID: 69fc539a47f34ce924df27dc.

[Reporting paper](<https://www.sota2.com/research/paper/69fc1cb45497b3b670c886b3.md>)

Recorded evaluation configuration: {"Backbone":"SD v1.5"}. This record does not assert independent reproduction.

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