# Class-Incremental Learning on ImageNet 500 base + 10x50 novel classes

Source: [Class-Incremental Learning on ImageNet 500 base + 10x50 novel classes](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-500-base-plus-10x50-novel-classes>)

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

Data updated: 2026-02-26T05:34:33.683Z

Evaluation of model performance under a Class Incremental Learning setting on a custom ImageNet split consisting of 500 base classes and 10 sequential tasks of 50 novel classes each.

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 500 base + 10x50 novel classes (incremental) |
| Metric | Accuracy (t=0..10) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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 (t=0..10)](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-500-base-plus-10x50-novel-classes.md?metric=Accuracy%20(t%3D0..10)>): Higher is better; unit: PERCENT.

- [Avg Incremental Accuracy](<https://www.sota2.com/research/sota/class-incremental-learning-on-imagenet-500-base-plus-10x50-novel-classes.md?metric=Avg%20Incremental%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (t=0..10) | Result date | Report paper |
| --- | --- | --- | --- |
| score fusion (best-Acc\_avg) | 70.67 | Not recorded | 68fa6d6c1e209131cea39c30 |
| score fusion (best-balanced) | 69.4 | Not recorded | 68fa6d6c1e209131cea39c30 |
| score fusion (best-Acc\_avg) | 66.23 | Not recorded | 68fa6d6c1e209131cea39c30 |
| score fusion (best-Acc\_all) | 65.66 | Not recorded | 68fa6d6c1e209131cea39c30 |
| score fusion (best-balanced) | 65.34 | Not recorded | 68fa6d6c1e209131cea39c30 |
| PODNet | 64.13 | Not recorded | 68fa6d6c1e209131cea39c30 |
| score fusion (best-Acc\_all) | 61.45 | Not recorded | 68fa6d6c1e209131cea39c30 |
| BiC | 44.31 | Not recorded | 68fa6d6c1e209131cea39c30 |

### score fusion (best-Acc\_avg)

Accuracy (t=0..10): 70.67. Result ID: 698fa1083a6c314849faa748.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Implementation](<https://github.com/amazon-research/sp-cil>)

Recorded evaluation configuration: {"Backbone":"ResNet18"}. This record does not assert independent reproduction.

### score fusion (best-balanced)

Accuracy (t=0..10): 69.4. Result ID: 698fa1083a6c314849faa747.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Implementation](<https://github.com/amazon-research/sp-cil>)

Recorded evaluation configuration: {"Backbone":"ResNet18"}. This record does not assert independent reproduction.

### score fusion (best-Acc\_avg)

Accuracy (t=0..10): 66.23. Result ID: 698fa1083a6c314849faa745.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Implementation](<https://github.com/amazon-research/sp-cil>)

Recorded evaluation configuration: {"Backbone":"ResNet10"}. This record does not assert independent reproduction.

### score fusion (best-Acc\_all)

Accuracy (t=0..10): 65.66. Result ID: 698fa1083a6c314849faa746.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Implementation](<https://github.com/amazon-research/sp-cil>)

Recorded evaluation configuration: {"Backbone":"ResNet18"}. This record does not assert independent reproduction.

### score fusion (best-balanced)

Accuracy (t=0..10): 65.34. Result ID: 698fa1083a6c314849faa744.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Implementation](<https://github.com/amazon-research/sp-cil>)

Recorded evaluation configuration: {"Backbone":"ResNet10"}. This record does not assert independent reproduction.

### PODNet

Accuracy (t=0..10): 64.13. Result ID: 698fa1083a6c314849faa742.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa686b1e209131cea1bffa.md>)

[Implementation](<https://www.sota2.com/github.com/arthurdouillard/incremental_learning.pytorch>)

Recorded evaluation configuration: {"Backbone":"ResNet18"}. This record does not assert independent reproduction.

### score fusion (best-Acc\_all)

Accuracy (t=0..10): 61.45. Result ID: 698fa1083a6c314849faa743.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

[Implementation](<https://github.com/amazon-research/sp-cil>)

Recorded evaluation configuration: {"Backbone":"ResNet10"}. This record does not assert independent reproduction.

### BiC

Accuracy (t=0..10): 44.31. Result ID: 698fa1083a6c314849faa741.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d6c1e209131cea39c30.md>)

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

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

Recorded evaluation configuration: {"Backbone":"ResNet18"}. This record does not assert independent reproduction.

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