# Subject-driven Stylized Image Generation on DreamBooth (StyleDrop/K-LoRA) (test)

Source: [Subject-driven Stylized Image Generation on DreamBooth (StyleDrop/K-LoRA) (test)](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test>)

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

Data updated: 2026-04-08T07:57:11.580Z

Evaluation of subject-driven stylized image generation models using the DreamBooth dataset, specifically incorporating StyleDrop and K-LoRA concepts/styles, assessed via Style Similarity, CLIP, and DINO metrics on the test split.

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 | Subject-driven Stylized Image Generation |
| Dataset (including any stated split) | DreamBooth (concepts) + StyleDrop and K-LoRA (styles) (test) |
| Metric | Style Similarity |
| Unit | PERCENT |
| 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

- [Style Similarity](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test.md?metric=Style%20Similarity>): Higher is better; unit: PERCENT.

- [CLIP Score](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test.md?metric=CLIP%20Score>): Higher is better; unit: PERCENT.

- [DINO Score](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test.md?metric=DINO%20Score>): Higher is better; unit: PERCENT.

- [Geometric Mean (Style, DINO)](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test.md?metric=Geometric%20Mean%20(Style%2C%20DINO)>): Higher is better; unit: PERCENT.

- [Geometric Mean (Style, CLIP)](<https://www.sota2.com/research/sota/subject-driven-stylized-image-generation-on-dreambooth-styledrop-k-lora-test.md?metric=Geometric%20Mean%20(Style%2C%20CLIP)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Style Similarity | Result date | Report paper |
| --- | --- | --- | --- |
| OrthoFuse | 61 | Not recorded | 69d5e1287d0172140ec75b9c |
| ZipLoRA | 49 | Not recorded | 69d5e1287d0172140ec75b9c |
| ZipLoRA | 49 | Not recorded | 69d5e1287d0172140ec75b9c |
| K-LoRA | 49 | Not recorded | 69d5e1287d0172140ec75b9c |
| Joint training | 48 | Not recorded | 69d5e1287d0172140ec75b9c |
| K-LoRA | 46 | Not recorded | 69d5e1287d0172140ec75b9c |

### OrthoFuse

Style Similarity: 61. Result ID: 69d60a1d47f34ce924be62f2.

[Reporting paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

Recorded evaluation configuration: {"Training approach":"Training-free","Merging time":"\< 1 sec"}. This record does not assert independent reproduction.

### ZipLoRA

Style Similarity: 49. Result ID: 69d60a1d47f34ce924be62ee.

[Reporting paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa713e1e209131cea50076.md>)

Recorded evaluation configuration: {"Training approach":"Training-based","Rank (r)":"8","Merging time":"4 minutes"}. This record does not assert independent reproduction.

### ZipLoRA

Style Similarity: 49. Result ID: 69d60a1d47f34ce924be62ef.

[Reporting paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa713e1e209131cea50076.md>)

Recorded evaluation configuration: {"Training approach":"Training-based","Rank (r)":"64","Merging time":"4 minutes"}. This record does not assert independent reproduction.

### K-LoRA

Style Similarity: 49. Result ID: 69d60a1d47f34ce924be62f1.

[Reporting paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72053.md>)

Recorded evaluation configuration: {"Training approach":"Training-free","Rank (r)":"64","Merging time":"\< 1 sec"}. This record does not assert independent reproduction.

### Joint training

Style Similarity: 48. Result ID: 69d60a1d47f34ce924be62ed.

[Reporting paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

Recorded evaluation configuration: {"Training approach":"Training-based","Merging time":"1.5 hours"}. This record does not assert independent reproduction.

### K-LoRA

Style Similarity: 46. Result ID: 69d60a1d47f34ce924be62f0.

[Reporting paper](<https://www.sota2.com/research/paper/69d5e1287d0172140ec75b9c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77411e209131cea72053.md>)

Recorded evaluation configuration: {"Training approach":"Training-free","Rank (r)":"8","Merging time":"\< 1 sec"}. This record does not assert independent reproduction.

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