# Image Representation on ImageNet-1K 256x256

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

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

Data updated: 2026-06-01T08:34:35.826Z

Evaluation of the quality of learned image representations on the ImageNet-1K dataset using inputs resized to 256x256 resolution, typically measured via linear probing accuracy.

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 Representation |
| Dataset (including any stated split) | ImageNet-1K 256x256 |
| Metric | Linear Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

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

## Recorded results

| Method | Linear Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| MergeTok-BL | 78.3 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| MergeTok-BL\* | 78.2 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| MergeVQ-GR | 77.9 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| MergeTok-SB | 73.8 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| UniTok\* | 70.8 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| VFMTok | 69.4 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| ViT-VQGAN | 65.1 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| GigaTok-BL\* | 64.1 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| GigaTok-BL | 63.8 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| GigaTok-SB | 61.5 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| Titok-L-32 | 60 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| Titok-B-64 | 53.9 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| LlamaGen | 47.6 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| Titok-S-128 | 46.6 | Not recorded | 6a1d11e1dbd7150506f83dd6 |
| Hita | 36.6 | Not recorded | 6a1d11e1dbd7150506f83dd6 |

### MergeTok-BL

Linear Accuracy: 78.3. Result ID: 6a1d4329fbc373592c1674d2.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VAE+VQ","VQ Codebook #Tok.":"256","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

### MergeTok-BL\*

Linear Accuracy: 78.2. Result ID: 6a1d4329fbc373592c1674d1.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VAE+VQ","VQ Codebook #Tok.":"256","VQ Codebook #Code":"2^14","DINO discriminator":"frozen"}. This record does not assert independent reproduction.

### MergeVQ-GR

Linear Accuracy: 77.9. Result ID: 6a1d4329fbc373592c1674cb.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77b41e209131cea74dd6.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D LFQ","VQ Codebook #Code":"2^18"}. This record does not assert independent reproduction.

### MergeTok-SB

Linear Accuracy: 73.8. Result ID: 6a1d4329fbc373592c1674d0.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VAE+VQ","VQ Codebook #Tok.":"256","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

### UniTok\*

Linear Accuracy: 70.8. Result ID: 6a1d4329fbc373592c1674c7.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78951e209131cea791bb.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"2D MSQ","VQ Codebook #Tok.":"16^2","VQ Codebook #Code":"2^15","DINO discriminator":"frozen"}. This record does not assert independent reproduction.

### VFMTok

Linear Accuracy: 69.4. Result ID: 6a1d4329fbc373592c1674c6.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79fb1e209131cea806ef.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"2D VQ","VQ Codebook #Tok.":"16^2","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

### ViT-VQGAN

Linear Accuracy: 65.1. Result ID: 6a1d4329fbc373592c1674c4.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cba1e209131cea35a1f.md>)

Recorded evaluation configuration: {"Ratio":"8","VQ Codebook Type":"2D VQ","VQ Codebook #Tok.":"32^2","VQ Codebook #Code":"2^13"}. This record does not assert independent reproduction.

### GigaTok-BL\*

Linear Accuracy: 64.1. Result ID: 6a1d4329fbc373592c1674cd.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea8530d.md>)

[Implementation](<https://silentview.github.io/GigaTok/>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"256","VQ Codebook #Code":"2^14","DINO discriminator":"frozen"}. This record does not assert independent reproduction.

### GigaTok-BL

Linear Accuracy: 63.8. Result ID: 6a1d4329fbc373592c1674ce.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea8530d.md>)

[Implementation](<https://silentview.github.io/GigaTok/>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"256","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

### GigaTok-SB

Linear Accuracy: 61.5. Result ID: 6a1d4329fbc373592c1674cc.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea8530d.md>)

[Implementation](<https://silentview.github.io/GigaTok/>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"256","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

### Titok-L-32

Linear Accuracy: 60. Result ID: 6a1d4329fbc373592c1674ca.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7e1.md>)

[Implementation](<https://yucornetto.github.io/projects/titok.html>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"32","VQ Codebook #Code":"2^12"}. This record does not assert independent reproduction.

### Titok-B-64

Linear Accuracy: 53.9. Result ID: 6a1d4329fbc373592c1674c9.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7e1.md>)

[Implementation](<https://yucornetto.github.io/projects/titok.html>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"64","VQ Codebook #Code":"2^12"}. This record does not assert independent reproduction.

### LlamaGen

Linear Accuracy: 47.6. Result ID: 6a1d4329fbc373592c1674c5.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738a1e209131cea5d111.md>)

[Implementation](<https://github.com/FoundationVision/LlamaGen>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"2D VQ","VQ Codebook #Tok.":"16^2","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

### Titok-S-128

Linear Accuracy: 46.6. Result ID: 6a1d4329fbc373592c1674c8.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d7e1.md>)

[Implementation](<https://yucornetto.github.io/projects/titok.html>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"128","VQ Codebook #Code":"2^12"}. This record does not assert independent reproduction.

### Hita

Linear Accuracy: 36.6. Result ID: 6a1d4329fbc373592c1674cf.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d11e1dbd7150506f83dd6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79f91e209131cea80635.md>)

Recorded evaluation configuration: {"Ratio":"16","VQ Codebook Type":"1D VQ","VQ Codebook #Tok.":"569","VQ Codebook #Code":"2^14"}. This record does not assert independent reproduction.

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