# Latent-guided Style Manipulation on CelebA-HQ

Source: [Latent-guided Style Manipulation on CelebA-HQ](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq>)

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

Data updated: 2026-04-24T07:31:39.026Z

Evaluation of latent-guided style manipulation capabilities on the CelebA-HQ dataset, focusing on attribute-specific FID and classification accuracy for targeted edits (e.g., Bangs, Glasses).

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 | Latent-guided Style Manipulation |
| Dataset (including any stated split) | CelebA-HQ |
| Metric | Bangs FID |
| Unit | SCALAR |
| Direction | Lower 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

- [Bangs FID](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Bangs%20FID>): Lower is better; unit: SCALAR.

- [Bangs Accuracy](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Bangs%20Accuracy>): Higher is better; unit: PERCENT.

- [Bangs Baseline FID](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Bangs%20Baseline%20FID>): Lower is better; unit: SCALAR.

- [Bangs Baseline Accuracy](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Bangs%20Baseline%20Accuracy>): Higher is better; unit: PERCENT.

- [Eleglasses FID](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Eleglasses%20FID>): Lower is better; unit: SCALAR.

- [Eleglasses Accuracy](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Eleglasses%20Accuracy>): Higher is better; unit: PERCENT.

- [Eleglasses Baseline FID](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Eleglasses%20Baseline%20FID>): Lower is better; unit: SCALAR.

- [Eleglasses Baseline Accuracy](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Eleglasses%20Baseline%20Accuracy>): Higher is better; unit: PERCENT.

- [Average FID](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Average%20FID>): Lower is better; unit: SCALAR.

- [Average Accuracy](<https://www.sota2.com/research/sota/latent-guided-style-manipulation-on-celeba-hq.md?metric=Average%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Bangs FID | Result date | Report paper |
| --- | --- | --- | --- |
| LatRef-Diff | 15.53 | Not recorded | 69eaf8f555a3c32af89f886b |
| SDGAN | 15.69 | Not recorded | 69eaf8f555a3c32af89f886b |
| BSGAN | 16.69 | Not recorded | 69eaf8f555a3c32af89f886b |
| VecGAN++ | 17.12 | Not recorded | 69eaf8f555a3c32af89f886b |
| HiSD | 17.18 | Not recorded | 69eaf8f555a3c32af89f886b |

### LatRef-Diff

Bangs FID: 15.53. Result ID: 69eb1c5047f34ce924d2f5e3.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf8f555a3c32af89f886b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69eaf8f555a3c32af89f886b.md>)

Recorded evaluation configuration: {"Setting":"#1","Guidance":"Latent-guided"}. This record does not assert independent reproduction.

### SDGAN

Bangs FID: 15.69. Result ID: 69eb1c5047f34ce924d2f5e1.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf8f555a3c32af89f886b.md>)

Recorded evaluation configuration: {"Setting":"#1","Guidance":"Latent-guided"}. This record does not assert independent reproduction.

### BSGAN

Bangs FID: 16.69. Result ID: 69eb1c5047f34ce924d2f5e2.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf8f555a3c32af89f886b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77b61e209131cea74eb6.md>)

Recorded evaluation configuration: {"Setting":"#1","Guidance":"Latent-guided"}. This record does not assert independent reproduction.

### VecGAN++

Bangs FID: 17.12. Result ID: 69eb1c5047f34ce924d2f5e0.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf8f555a3c32af89f886b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77b61e209131cea74eb6.md>)

Recorded evaluation configuration: {"Setting":"#1","Guidance":"Latent-guided"}. This record does not assert independent reproduction.

### HiSD

Bangs FID: 17.18. Result ID: 69eb1c5047f34ce924d2f5df.

[Reporting paper](<https://www.sota2.com/research/paper/69eaf8f555a3c32af89f886b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69741e209131cea222ee.md>)

[Implementation](<https://github.com/imlixinyang/HiSD>)

Recorded evaluation configuration: {"Setting":"#1","Guidance":"Latent-guided"}. This record does not assert independent reproduction.

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