# Image Classification on SVHN Split

Source: [Image Classification on SVHN Split](<https://www.sota2.com/research/sota/image-classification-on-svhn-split>)

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

Data updated: 2026-02-26T05:34:28.030Z

Evaluation of Image Classification performance on the specific 'Split' configuration of the SVHN dataset, reporting accuracy across five distinct test conditions (T1-T5) and the average.

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) | SVHN Split |
| Metric | Accuracy (T1) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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 (T1)](<https://www.sota2.com/research/sota/image-classification-on-svhn-split.md?metric=Accuracy%20(T1)>): Higher is better; unit: PERCENT.

- [Accuracy (T2)](<https://www.sota2.com/research/sota/image-classification-on-svhn-split.md?metric=Accuracy%20(T2)>): Higher is better; unit: PERCENT.

- [Accuracy (T3)](<https://www.sota2.com/research/sota/image-classification-on-svhn-split.md?metric=Accuracy%20(T3)>): Higher is better; unit: PERCENT.

- [Accuracy (T4)](<https://www.sota2.com/research/sota/image-classification-on-svhn-split.md?metric=Accuracy%20(T4)>): Higher is better; unit: PERCENT.

- [Accuracy (T5)](<https://www.sota2.com/research/sota/image-classification-on-svhn-split.md?metric=Accuracy%20(T5)>): Higher is better; unit: PERCENT.

- [Average Accuracy](<https://www.sota2.com/research/sota/image-classification-on-svhn-split.md?metric=Average%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (T1) | Result date | Report paper |
| --- | --- | --- | --- |
| Joint (UB) | 98.3 | Not recorded | 68fa67551e209131cea1592c |
| Conditional Channel Gated Networks | 97.8 | Not recorded | 68fa67551e209131cea1592c |
| LwF | 97.4 | Not recorded | 68fa67551e209131cea1592c |
| HAT | 97.1 | Not recorded | 68fa67551e209131cea1592c |
| EWC-On | 96.6 | Not recorded | 68fa67551e209131cea1592c |

### Joint (UB)

Accuracy (T1): 98.3. Result ID: 698d2c818b517eb4331f8de9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67551e209131cea1592c.md>)

Recorded evaluation configuration: {"Setting":"Task-incremental","Role":"Upper Bound"}. This record does not assert independent reproduction.

### Conditional Channel Gated Networks

Accuracy (T1): 97.8. Result ID: 698d2c818b517eb4331f8ded.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67551e209131cea1592c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67551e209131cea1592c.md>)

Recorded evaluation configuration: {"Setting":"Task-incremental"}. This record does not assert independent reproduction.

### LwF

Accuracy (T1): 97.4. Result ID: 698d2c818b517eb4331f8deb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67551e209131cea1592c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63201e209131ce9fd0b3.md>)

Recorded evaluation configuration: {"Setting":"Task-incremental"}. This record does not assert independent reproduction.

### HAT

Accuracy (T1): 97.1. Result ID: 698d2c818b517eb4331f8dec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67551e209131cea1592c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64681e209131cea0485c.md>)

[Implementation](<https://github.com/joansj/hat>)

Recorded evaluation configuration: {"Setting":"Task-incremental"}. This record does not assert independent reproduction.

### EWC-On

Accuracy (T1): 96.6. Result ID: 698d2c818b517eb4331f8dea.

[Reporting paper](<https://www.sota2.com/research/paper/68fa67551e209131cea1592c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64881e209131cea053e1.md>)

Recorded evaluation configuration: {"Setting":"Task-incremental"}. This record does not assert independent reproduction.

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