# Counterfactual Image Generation on MS-COCO subset (test)

Source: [Counterfactual Image Generation on MS-COCO subset (test)](<https://www.sota2.com/research/sota/counterfactual-image-generation-on-ms-coco-subset-test>)

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

Data updated: 2026-05-25T09:09:40.209Z

Evaluation of models performing counterfactual image generation on a specific subset of the MS-COCO test set, utilizing metrics like VLM-Eff. and LPIPS.

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 | Counterfactual image generation |
| Dataset (including any stated split) | MS-COCO subset (test) |
| Metric | VLM-Eff |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [VLM-Eff](<https://www.sota2.com/research/sota/counterfactual-image-generation-on-ms-coco-subset-test.md?metric=VLM-Eff>): Higher is better; unit: SCALAR.

- [LPIPS](<https://www.sota2.com/research/sota/counterfactual-image-generation-on-ms-coco-subset-test.md?metric=LPIPS>): Lower is better; unit: SCALAR.

## Recorded results

| Method | VLM-Eff | Result date | Report paper |
| --- | --- | --- | --- |
| CSG | 0.762 | Not recorded | 6a13d7ad6f71c5ae7dc9eab9 |
| DDIM Inversion | 0.741 | Not recorded | 6a13d7ad6f71c5ae7dc9eab9 |
| DDPM Inversion | 0.655 | Not recorded | 6a13d7ad6f71c5ae7dc9eab9 |

### CSG

VLM-Eff: 0.762. Result ID: 6a1411bafbc373592c0b7e52.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7ad6f71c5ae7dc9eab9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a13d7ad6f71c5ae7dc9eab9.md>)

Recorded evaluation configuration: {"Backbone":"Stable Diffusion XL"}. This record does not assert independent reproduction.

### DDIM Inversion

VLM-Eff: 0.741. Result ID: 6a1411bafbc373592c0b7e50.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7ad6f71c5ae7dc9eab9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8b1e209131cea3a7fe.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### DDPM Inversion

VLM-Eff: 0.655. Result ID: 6a1411bafbc373592c0b7e51.

[Reporting paper](<https://www.sota2.com/research/paper/6a13d7ad6f71c5ae7dc9eab9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68ef1e209131cea1f0d5.md>)

[Implementation](<https://github.com/hojonathanho/diffusion>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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