# Image Inpainting on ImageNet 1k 256x256

Source: [Image Inpainting on ImageNet 1k 256x256](<https://www.sota2.com/research/sota/image-inpainting-on-imagenet-1k-256x256>)

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

Data updated: 2026-07-03T06:49:31.447Z

Evaluation of image inpainting performance on the ImageNet 1k dataset using 256x256 resolution inputs, measured using FID, IS, Precision (P), and Recall (R).

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 Inpainting |
| Dataset (including any stated split) | ImageNet 1k 256x256 |
| Metric | FID |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 7 of 7 |
| 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

- [FID](<https://www.sota2.com/research/sota/image-inpainting-on-imagenet-1k-256x256.md?metric=FID>): Lower is better; unit: SCALAR.

- [Inception Score](<https://www.sota2.com/research/sota/image-inpainting-on-imagenet-1k-256x256.md?metric=Inception%20Score>): Higher is better; unit: SCALAR.

- [Precision](<https://www.sota2.com/research/sota/image-inpainting-on-imagenet-1k-256x256.md?metric=Precision>): Higher is better; unit: PERCENT.

- [Recall](<https://www.sota2.com/research/sota/image-inpainting-on-imagenet-1k-256x256.md?metric=Recall>): Higher is better; unit: PERCENT.

## Recorded results

| Method | FID | Result date | Report paper |
| --- | --- | --- | --- |
| STAR-XL | 2.07 | Not recorded | 6923e65cd6b0dd334292ed2f |
| STAR-L | 2.2 | Not recorded | 6923e65cd6b0dd334292ed2f |
| RandAR-XL | 2.58 | Not recorded | 6923e65cd6b0dd334292ed2f |
| RandAR-L | 2.66 | Not recorded | 6923e65cd6b0dd334292ed2f |
| RAR-XL | 3.4 | Not recorded | 6923e65cd6b0dd334292ed2f |
| RAR-L | 3.56 | Not recorded | 6923e65cd6b0dd334292ed2f |
| DiT-XL/2 | 4.58 | Not recorded | 6923e65cd6b0dd334292ed2f |

### STAR-XL

FID: 2.07. Result ID: 6a475b7b60b88bb339922ead.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

Recorded evaluation configuration: {"Model Type":"Randomized AR","# Params":"955M"}. This record does not assert independent reproduction.

### STAR-L

FID: 2.2. Result ID: 6a475b7b60b88bb339922eac.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

Recorded evaluation configuration: {"Model Type":"Randomized AR","# Params":"461M"}. This record does not assert independent reproduction.

### RandAR-XL

FID: 2.58. Result ID: 6a475b7b60b88bb339922eab.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79ef1e209131cea80181.md>)

Recorded evaluation configuration: {"Model Type":"Randomized AR","# Params":"775M"}. This record does not assert independent reproduction.

### RandAR-L

FID: 2.66. Result ID: 6a475b7b60b88bb339922eaa.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79ef1e209131cea80181.md>)

Recorded evaluation configuration: {"Model Type":"Randomized AR","# Params":"343M"}. This record does not assert independent reproduction.

### RAR-XL

FID: 3.4. Result ID: 6a475b7b60b88bb339922ea9.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa758e1e209131cea6896a.md>)

[Implementation](<https://github.com/bytedance/1d-tokenizer>)

Recorded evaluation configuration: {"Model Type":"Raster-scan AR","# Params":"955M"}. This record does not assert independent reproduction.

### RAR-L

FID: 3.56. Result ID: 6a475b7b60b88bb339922ea8.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa758e1e209131cea6896a.md>)

[Implementation](<https://github.com/bytedance/1d-tokenizer>)

Recorded evaluation configuration: {"Model Type":"Raster-scan AR","# Params":"461M"}. This record does not assert independent reproduction.

### DiT-XL/2

FID: 4.58. Result ID: 6a475b7b60b88bb339922ea7.

[Reporting paper](<https://www.sota2.com/research/paper/6923e65cd6b0dd334292ed2f.md>)

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

Recorded evaluation configuration: {"Model Type":"Diffusion","# Params":"675M"}. This record does not assert independent reproduction.

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