# Continual Learning Accuracy on ImageNet-A

Source: [Continual Learning Accuracy on ImageNet-A](<https://www.sota2.com/research/sota/continual-learning-accuracy-on-imagenet-a>)

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

Data updated: 2026-02-27T09:48:39.879Z

Evaluation of model accuracy when performing Continual Learning tasks on the ImageNet-A 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 | Continual Learning |
| Dataset (including any stated split) | ImageNet-A |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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/continual-learning-accuracy-on-imagenet-a.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Joint | 70.8 | Not recorded | 68fa75d81e209131cea6a32b |
| RanDumb | 42.2 | Not recorded | 68fa75d81e209131cea6a32b |
| RanPAC-imp | 33.8 | Not recorded | 68fa75d81e209131cea6a32b |

### Joint

Accuracy: 70.8. Result ID: 69a1686835dfcabb00e6df2f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75d81e209131cea6a32b.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B16","Pre-training":"ImageNet21K","Evaluation Protocol":"Offline","Initial Classes":"2","New Classes per Task":"1"}. This record does not assert independent reproduction.

### RanDumb

Accuracy: 42.2. Result ID: 69a1686835dfcabb00e6df2e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75d81e209131cea6a32b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d81e209131cea6a32b.md>)

[Implementation](<https://github.com/drimpossible/RanDumb>)

Recorded evaluation configuration: {"Backbone":"ViT-B16","Pre-training":"ImageNet21K","Evaluation Protocol":"Offline","Initial Classes":"2","New Classes per Task":"1"}. This record does not assert independent reproduction.

### RanPAC-imp

Accuracy: 33.8. Result ID: 69a1686835dfcabb00e6df2d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75d81e209131cea6a32b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71cb1e209131cea53052.md>)

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

Recorded evaluation configuration: {"Backbone":"ViT-B16","Pre-training":"ImageNet21K","Evaluation Protocol":"Offline","Initial Classes":"2","New Classes per Task":"1"}. This record does not assert independent reproduction.

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