# Image Classification on 64x64 images

Source: [Image Classification on 64x64 images](<https://www.sota2.com/research/sota/image-classification-on-64x64-images>)

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

Data updated: 2026-02-26T05:48:28.566Z

Evaluation of Image Classification performance on a dataset composed of 64x64 resolution images, focusing on inference speed metrics like latency and throughput.

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 Classification |
| Dataset (including any stated split) | 64x64 images |
| Metric | Latency (ms) |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| 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

- [Latency (ms)](<https://www.sota2.com/research/sota/image-classification-on-64x64-images.md?metric=Latency%20(ms)>): Lower is better; unit: TIME.

- [Throughput (img/s)](<https://www.sota2.com/research/sota/image-classification-on-64x64-images.md?metric=Throughput%20(img%2Fs)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Latency (ms) | Result date | Report paper |
| --- | --- | --- | --- |
| B-12/Wi-768 | 21.2 | Not recorded | 68fa70b81e209131cea4ccd1 |
| Mixer-B/4 | 400 | Not recorded | 68fa70b81e209131cea4ccd1 |
| ResNet-152 | 423 | Not recorded | 68fa70b81e209131cea4ccd1 |
| ViT-B/4 | 424 | Not recorded | 68fa70b81e209131cea4ccd1 |

### B-12/Wi-768

Latency (ms): 21.2. Result ID: 699155723a6c314849076239.

[Reporting paper](<https://www.sota2.com/research/paper/68fa70b81e209131cea4ccd1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70b81e209131cea4ccd1.md>)

[Implementation](<https://github.com/gregorbachmann/scaling_mlps>)

Recorded evaluation configuration: {"Parameters":"66.89 M","FLOPs per forward pass":"66.8 M"}. This record does not assert independent reproduction.

### Mixer-B/4

Latency (ms): 400. Result ID: 699155723a6c31484907623c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa70b81e209131cea4ccd1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a201e209131cea26207.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Parameters":"63.82 M","FLOPs per forward pass":"19.36 G"}. This record does not assert independent reproduction.

### ResNet-152

Latency (ms): 423. Result ID: 699155723a6c31484907623a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa70b81e209131cea4ccd1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62321e209131ce9f8328.md>)

Recorded evaluation configuration: {"Parameters":"60.19 M","FLOPs per forward pass":"13.07 G"}. This record does not assert independent reproduction.

### ViT-B/4

Latency (ms): 424. Result ID: 699155723a6c31484907623b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa70b81e209131cea4ccd1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0b1e209131cea25b04.md>)

[Implementation](<https://github.com/google-research/vision_transformer>)

Recorded evaluation configuration: {"Parameters":"86.06 M","FLOPs per forward pass":"23.08 G"}. This record does not assert independent reproduction.

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