# Class-incremental Learning on ImageNet-100 (B=5, C=5)

Source: [Class-incremental Learning on ImageNet-100 (B=5, C=5)](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-100-b-5-c-5>)

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

Data updated: 2026-02-27T09:02:31.590Z

Evaluation of Class-Incremental Learning performance on the specific ImageNet-100 subset defined by B=5 classes and C=5 tasks/steps, reporting Average incremental accuracy, Last-step accuracy, and Forgetting rate.

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 | Class-incremental Learning |
| Dataset (including any stated split) | ImageNet-100 B=5, C=5 1.0 |
| Metric | Average Incremental Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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

- [Average Incremental Accuracy](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-100-b-5-c-5.md?metric=Average%20Incremental%20Accuracy>): Higher is better; unit: PERCENT.

- [Last-Step Accuracy](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-100-b-5-c-5.md?metric=Last-Step%20Accuracy>): Higher is better; unit: PERCENT.

- [Forgetting Rate](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-100-b-5-c-5.md?metric=Forgetting%20Rate>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Average Incremental Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| DyTox w/ LingoCL | 76.2 | Not recorded | 68fa729c1e209131cea577e0 |
| DyTox | 75.4 | Not recorded | 68fa729c1e209131cea577e0 |
| BiC w/ LingoCL | 61.9 | Not recorded | 68fa729c1e209131cea577e0 |
| BiC | 59 | Not recorded | 68fa729c1e209131cea577e0 |
| LUCIR w/ LingoCL | 53.6 | Not recorded | 68fa729c1e209131cea577e0 |
| IL2M w/ LingoCL | 51.7 | Not recorded | 68fa729c1e209131cea577e0 |
| AANet w/ LingoCL | 50.1 | Not recorded | 68fa729c1e209131cea577e0 |
| BiMeCo w/ LingoCL | 49.8 | Not recorded | 68fa729c1e209131cea577e0 |
| AANet | 48.5 | Not recorded | 68fa729c1e209131cea577e0 |
| BiMeCo | 47.6 | Not recorded | 68fa729c1e209131cea577e0 |
| LUCIR | 47.5 | Not recorded | 68fa729c1e209131cea577e0 |
| IL2M | 46.4 | Not recorded | 68fa729c1e209131cea577e0 |
| CwD w/ LingoCL | 41 | Not recorded | 68fa729c1e209131cea577e0 |
| CwD | 38.9 | Not recorded | 68fa729c1e209131cea577e0 |

### DyTox w/ LingoCL

Average Incremental Accuracy: 76.2. Result ID: 69a15da235dfcabb00e69084.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Architecture-based"}. This record does not assert independent reproduction.

### DyTox

Average Incremental Accuracy: 75.4. Result ID: 69a15da235dfcabb00e69083.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d351e209131cea386d9.md>)

[Implementation](<https://github.com/arthurdouillard/dytox>)

Recorded evaluation configuration: {"Method Category":"Architecture-based"}. This record does not assert independent reproduction.

### BiC w/ LingoCL

Average Incremental Accuracy: 61.9. Result ID: 69a15da235dfcabb00e6908a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Distillation-based"}. This record does not assert independent reproduction.

### BiC

Average Incremental Accuracy: 59. Result ID: 69a15da235dfcabb00e69089.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65dc1e209131cea0ce26.md>)

[Implementation](<https://github.com/srebuffi/iCaRL>)

Recorded evaluation configuration: {"Method Category":"Distillation-based"}. This record does not assert independent reproduction.

### LUCIR w/ LingoCL

Average Incremental Accuracy: 53.6. Result ID: 69a15da235dfcabb00e69088.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Distillation-based"}. This record does not assert independent reproduction.

### IL2M w/ LingoCL

Average Incremental Accuracy: 51.7. Result ID: 69a15da235dfcabb00e6908e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Rectification-based"}. This record does not assert independent reproduction.

### AANet w/ LingoCL

Average Incremental Accuracy: 50.1. Result ID: 69a15da235dfcabb00e69082.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Architecture-based"}. This record does not assert independent reproduction.

### BiMeCo w/ LingoCL

Average Incremental Accuracy: 49.8. Result ID: 69a15da235dfcabb00e69086.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Architecture-based"}. This record does not assert independent reproduction.

### AANet

Average Incremental Accuracy: 48.5. Result ID: 69a15da235dfcabb00e69081.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bd61e209131cea30751.md>)

Recorded evaluation configuration: {"Method Category":"Architecture-based"}. This record does not assert independent reproduction.

### BiMeCo

Average Incremental Accuracy: 47.6. Result ID: 69a15da235dfcabb00e69085.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Architecture-based"}. This record does not assert independent reproduction.

### LUCIR

Average Incremental Accuracy: 47.5. Result ID: 69a15da235dfcabb00e69087.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Distillation-based"}. This record does not assert independent reproduction.

### IL2M

Average Incremental Accuracy: 46.4. Result ID: 69a15da235dfcabb00e6908d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Rectification-based"}. This record does not assert independent reproduction.

### CwD w/ LingoCL

Average Incremental Accuracy: 41. Result ID: 69a15da235dfcabb00e6908c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

Recorded evaluation configuration: {"Method Category":"Rectification-based"}. This record does not assert independent reproduction.

### CwD

Average Incremental Accuracy: 38.9. Result ID: 69a15da235dfcabb00e6908b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729c1e209131cea577e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72d51e209131cea58ce9.md>)

[Implementation](<https://github.com/Yujun-Shi/CwD>)

Recorded evaluation configuration: {"Method Category":"Rectification-based"}. This record does not assert independent reproduction.

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