# Multi-channel Image Classification on CHAMMI HPA (test)

Source: [Multi-channel Image Classification on CHAMMI HPA (test)](<https://www.sota2.com/research/sota/multi-channel-image-classification-on-chammi-hpa-test>)

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

Data updated: 2026-02-26T05:30:00.586Z

Evaluation of Multi-channel Image Classification performance on the test split of the CHAMMI HPA dataset.

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 | Multi-channel Image Classification |
| Dataset (including any stated split) | CHAMMI HPA (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher 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

- [Accuracy](<https://www.sota2.com/research/sota/multi-channel-image-classification-on-chammi-hpa-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Depthwise | 71.52 | Not recorded | 68fa756e1e209131cea67df2 |
| HyperNet | 65.93 | Not recorded | 68fa756e1e209131cea67df2 |
| TemplateMixing | 64.52 | Not recorded | 68fa756e1e209131cea67df2 |
| HyperNet | 63.9 | Not recorded | 68fa756e1e209131cea67df2 |
| DiChaViT | 63.67 | Not recorded | 68fa756e1e209131cea67df2 |
| TemplateMixing | 63.44 | Not recorded | 68fa756e1e209131cea67df2 |
| ChannelViT | 62.14 | Not recorded | 68fa756e1e209131cea67df2 |
| ChAda-ViT | 60.67 | Not recorded | 68fa756e1e209131cea67df2 |
| Depthwise | 57.6 | Not recorded | 68fa756e1e209131cea67df2 |

### Depthwise

Accuracy: 71.52. Result ID: 6993a3343a6c314849177318.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71cd1e209131cea5309b.md>)

[Implementation](<https://github.com/chaudatascience/channel_adaptive_models>)

Recorded evaluation configuration: {"Architecture":"ViT"}. This record does not assert independent reproduction.

### HyperNet

Accuracy: 65.93. Result ID: 6993a3343a6c314849177314.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62fd1e209131ce9fc06b.md>)

Recorded evaluation configuration: {"Architecture":"ConvNeXt"}. This record does not assert independent reproduction.

### TemplateMixing

Accuracy: 64.52. Result ID: 6993a3343a6c314849177319.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfa1e209131cea31605.md>)

Recorded evaluation configuration: {"Architecture":"ViT"}. This record does not assert independent reproduction.

### HyperNet

Accuracy: 63.9. Result ID: 6993a3343a6c314849177317.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62fd1e209131ce9fc06b.md>)

Recorded evaluation configuration: {"Architecture":"ViT"}. This record does not assert independent reproduction.

### DiChaViT

Accuracy: 63.67. Result ID: 6993a3343a6c31484917731c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Implementation](<https://github.com/chaudatascience/diverse_channel_vit>)

Recorded evaluation configuration: {"Architecture":"ViT"}. This record does not assert independent reproduction.

### TemplateMixing

Accuracy: 63.44. Result ID: 6993a3343a6c314849177316.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bfa1e209131cea31605.md>)

Recorded evaluation configuration: {"Architecture":"ConvNeXt"}. This record does not assert independent reproduction.

### ChannelViT

Accuracy: 62.14. Result ID: 6993a3343a6c31484917731b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72ed1e209131cea594b5.md>)

Recorded evaluation configuration: {"Architecture":"ViT"}. This record does not assert independent reproduction.

### ChAda-ViT

Accuracy: 60.67. Result ID: 6993a3343a6c31484917731a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73651e209131cea5c2e7.md>)

Recorded evaluation configuration: {"Architecture":"ViT"}. This record does not assert independent reproduction.

### Depthwise

Accuracy: 57.6. Result ID: 6993a3343a6c314849177315.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756e1e209131cea67df2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71cd1e209131cea5309b.md>)

[Implementation](<https://github.com/chaudatascience/channel_adaptive_models>)

Recorded evaluation configuration: {"Architecture":"ConvNeXt"}. This record does not assert independent reproduction.

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