# Image Generation on ImageNet 256x256 (KD, FD, FLS Metrics)

Source: [Image Generation on ImageNet 256x256 (KD, FD, FLS Metrics)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-256x256-kd-fd-fls-metrics>)

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

Data updated: 2026-06-26T07:48:42.878Z

Evaluation of image generation models on ImageNet 256x256, reporting metrics based on Kernel Distance (KD), Fréchet Distance (FD), and Fréchet L2 Score (FLS).

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 | Image Generation |
| Dataset (including any stated split) | ImageNet 256x256 |
| Metric | KD (All) |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 9 of 9 |
| 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

- [KD (All)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-256x256-kd-fd-fls-metrics.md?metric=KD%20(All)>): Lower is better; unit: SCALAR.

- [FD (All)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-256x256-kd-fd-fls-metrics.md?metric=FD%20(All)>): Lower is better; unit: SCALAR.

- [FD-inf (All)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-256x256-kd-fd-fls-metrics.md?metric=FD-inf%20(All)>): Lower is better; unit: SCALAR.

- [FLS (All)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-256x256-kd-fd-fls-metrics.md?metric=FLS%20(All)>): Higher is better; unit: SCALAR.

- [KD (Avg)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-256x256-kd-fd-fls-metrics.md?metric=KD%20(Avg)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | KD (All) | Result date | Report paper |
| --- | --- | --- | --- |
| DiT-XL-2 | 0.06 | Not recorded | 68fa73991e209131cea5d7ed |
| LDM | 0.09 | Not recorded | 68fa73991e209131cea5d7ed |
| ADM | 0.19 | Not recorded | 68fa73991e209131cea5d7ed |
| StyleGAN-XL | 0.22 | Not recorded | 68fa73991e209131cea5d7ed |
| GigaGAN | 0.24 | Not recorded | 68fa73991e209131cea5d7ed |
| Config-A | 0.25 | Not recorded | 68fa73991e209131cea5d7ed |
| Config-B | 0.3 | Not recorded | 68fa73991e209131cea5d7ed |
| Config-C | 0.36 | Not recorded | 68fa73991e209131cea5d7ed |
| BigGAN | 0.52 | Not recorded | 68fa73991e209131cea5d7ed |

### DiT-XL-2

KD (All): 0.06. Result ID: 6a3e2ec1fbc373592c5f8879.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ea91e209131cea41328.md>)

Recorded evaluation configuration: {"Train Setup":"Centralized"}. This record does not assert independent reproduction.

### LDM

KD (All): 0.09. Result ID: 6a3e2ec1fbc373592c5f8878.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

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

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

Recorded evaluation configuration: {"Train Setup":"Centralized"}. This record does not assert independent reproduction.

### ADM

KD (All): 0.19. Result ID: 6a3e2ec1fbc373592c5f8877.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0e1e209131cea25dba.md>)

[Implementation](<https://github.com/openai/guided-diffusion>)

Recorded evaluation configuration: {"Train Setup":"Centralized"}. This record does not assert independent reproduction.

### StyleGAN-XL

KD (All): 0.22. Result ID: 6a3e2ec1fbc373592c5f8876.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c741e209131cea34012.md>)

Recorded evaluation configuration: {"Train Setup":"Centralized"}. This record does not assert independent reproduction.

### GigaGAN

KD (All): 0.24. Result ID: 6a3e2ec1fbc373592c5f8875.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fcb1e209131cea4775d.md>)

Recorded evaluation configuration: {"Train Setup":"Centralized"}. This record does not assert independent reproduction.

### Config-A

KD (All): 0.25. Result ID: 6a3e2ec1fbc373592c5f887c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

Recorded evaluation configuration: {"Train Setup":"Distributed"}. This record does not assert independent reproduction.

### Config-B

KD (All): 0.3. Result ID: 6a3e2ec1fbc373592c5f887b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

Recorded evaluation configuration: {"Train Setup":"Distributed"}. This record does not assert independent reproduction.

### Config-C

KD (All): 0.36. Result ID: 6a3e2ec1fbc373592c5f887a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

Recorded evaluation configuration: {"Train Setup":"Distributed"}. This record does not assert independent reproduction.

### BigGAN

KD (All): 0.52. Result ID: 6a3e2ec1fbc373592c5f8874.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7ed.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65661e209131cea0a090.md>)

[Implementation](<https://tfhub.dev/s?q=biggan>)

Recorded evaluation configuration: {"Train Setup":"Centralized"}. This record does not assert independent reproduction.

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