# Image Generation on ImageNet 128x128 25% 1.0 (train)

Source: [Image Generation on ImageNet 128x128 25% 1.0 (train)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-128x128-25-percent-1-0-train>)

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

Data updated: 2026-02-26T05:47:05.622Z

Evaluation of image generation models using a 25% subset of the training split (version 1.0) of the ImageNet dataset rendered at 128x128 resolution, measured by Inception Score (IS) and Fréchet Inception Distance (FID).

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 128x128 25% 1.0 (train) |
| Metric | Inception Score (IS) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [Inception Score (IS)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-128x128-25-percent-1-0-train.md?metric=Inception%20Score%20(IS)>): Higher is better; unit: SCALAR.

- [Fréchet Inception Distance (FID)](<https://www.sota2.com/research/sota/image-generation-on-imagenet-128x128-25-percent-1-0-train.md?metric=Fr%C3%A9chet%20Inception%20Distance%20(FID)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Inception Score (IS) | Result date | Report paper |
| --- | --- | --- | --- |
| BigGAN + DiffAugment | 74.2 | Not recorded | 68fa68df1e209131cea1edba |
| BigGAN | 46.5 | Not recorded | 68fa68df1e209131cea1edba |

### BigGAN + DiffAugment

Inception Score (IS): 74.2. Result ID: 698d8b728b517eb433224651.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68df1e209131cea1edba.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68df1e209131cea1edba.md>)

[Implementation](<https://github.com/mit-han-lab/data-efficient-gans>)

Recorded evaluation configuration: {"Truncation trick":"false","Number of evaluation samples":"50k","Number of evaluation runs":"3"}. This record does not assert independent reproduction.

### BigGAN

Inception Score (IS): 46.5. Result ID: 698d8b728b517eb433224650.

[Reporting paper](<https://www.sota2.com/research/paper/68fa68df1e209131cea1edba.md>)

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

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

Recorded evaluation configuration: {"Truncation trick":"false","Number of evaluation samples":"50k","Number of evaluation runs":"3"}. This record does not assert independent reproduction.

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