# Continual Learning on {MNIST, SVHN, notMNIST, FashionMNIST, CIFAR-10} (test)

Source: [Continual Learning on {MNIST, SVHN, notMNIST, FashionMNIST, CIFAR-10} (test)](<https://www.sota2.com/research/sota/continual-learning-on-mnist-svhn-notmnist-fashionmnist-cifar-10-test>)

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

Data updated: 2026-05-19T08:05:53.615Z

Evaluation of Continual Learning performance across the combined set of five datasets: MNIST, SVHN, notMNIST, FashionMNIST, and CIFAR-10 (test split), reporting AA and AF metrics.

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) | 5-dataset MNIST, SVHN, notMNIST, FashionMNIST, CIFAR-10 (test) |
| Metric | AA |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 17 of 17 |
| 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

- [AA](<https://www.sota2.com/research/sota/continual-learning-on-mnist-svhn-notmnist-fashionmnist-cifar-10-test.md?metric=AA>): Higher is better; unit: PERCENT.

- [AF](<https://www.sota2.com/research/sota/continual-learning-on-mnist-svhn-notmnist-fashionmnist-cifar-10-test.md?metric=AF>): Lower is better; unit: PERCENT.

- [FGT](<https://www.sota2.com/research/sota/continual-learning-on-mnist-svhn-notmnist-fashionmnist-cifar-10-test.md?metric=FGT>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AA | Result date | Report paper |
| --- | --- | --- | --- |
| DECODE | 90.56 | Not recorded | 68fa75e81e209131cea6a909 |
| Dualprompt | 88.08 | Not recorded | 68fa75e81e209131cea6a909 |
| PEFT Ensemble (FiLM) | 87.83 | Not recorded | 68fa78bb1e209131cea7a1ca |
| DER++ | 84.88 | Not recorded | 68fa75e81e209131cea6a909 |
| Coda-P | 84.82 | Not recorded | 68fa75e81e209131cea6a909 |
| ER | 84.26 | Not recorded | 68fa75e81e209131cea6a909 |
| F2P | 81.14 | Not recorded | 68fa75e81e209131cea6a909 |
| PEFT Ensemble (FiLM) | 79.69 | Not recorded | 68fa78bb1e209131cea7a1ca |
| PEFT Ensemble (FiLM) | 65.75 | Not recorded | 68fa78bb1e209131cea7a1ca |
| Cosine classifier | 59.87 | Not recorded | 68fa78bb1e209131cea7a1ca |
| Cosine classifier | 59.78 | Not recorded | 68fa78bb1e209131cea7a1ca |
| Cosine classifier | 58.54 | Not recorded | 68fa78bb1e209131cea7a1ca |
| EWC | 50.93 | Not recorded | 68fa75e81e209131cea6a909 |
| LwF | 47.91 | Not recorded | 68fa75e81e209131cea6a909 |
| Naive | 17.56 | Not recorded | 68fa78bb1e209131cea7a1ca |
| Naive | 15.93 | Not recorded | 68fa78bb1e209131cea7a1ca |
| Naive | 13.15 | Not recorded | 68fa78bb1e209131cea7a1ca |

### DECODE

AA: 90.56. Result ID: 6a0c19dffbc373592c0209a0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

Recorded evaluation configuration: {"Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### Dualprompt

AA: 88.08. Result ID: 6a0c19dffbc373592c02099e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d341e209131cea38614.md>)

Recorded evaluation configuration: {"Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### PEFT Ensemble (FiLM)

AA: 87.83. Result ID: 69ef008847f34ce924d470aa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"8","Privacy parameter (delta)":"1e-5","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### DER++

AA: 84.88. Result ID: 6a0c19dffbc373592c02099a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d341e209131cea38614.md>)

Recorded evaluation configuration: {"Memory Buffer Size":"500","Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### Coda-P

AA: 84.82. Result ID: 6a0c19dffbc373592c02099f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ee41e209131cea42add.md>)

Recorded evaluation configuration: {"Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### ER

AA: 84.26. Result ID: 6a0c19dffbc373592c020999.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d341e209131cea38614.md>)

Recorded evaluation configuration: {"Memory Buffer Size":"500","Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### F2P

AA: 81.14. Result ID: 6a0c19dffbc373592c02099d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d341e209131cea38614.md>)

Recorded evaluation configuration: {"Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### PEFT Ensemble (FiLM)

AA: 79.69. Result ID: 69ef008847f34ce924d470a9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"1","Privacy parameter (delta)":"1e-5","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### PEFT Ensemble (FiLM)

AA: 65.75. Result ID: 69ef008847f34ce924d470ab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"non-DP","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### Cosine classifier

AA: 59.87. Result ID: 69ef008847f34ce924d470a8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"non-DP","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### Cosine classifier

AA: 59.78. Result ID: 69ef008847f34ce924d470a7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"8","Privacy parameter (delta)":"1e-5","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### Cosine classifier

AA: 58.54. Result ID: 69ef008847f34ce924d470a6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"1","Privacy parameter (delta)":"1e-5","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### EWC

AA: 50.93. Result ID: 6a0c19dffbc373592c02099b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d341e209131cea38614.md>)

Recorded evaluation configuration: {"Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### LwF

AA: 47.91. Result ID: 6a0c19dffbc373592c02099c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa75e81e209131cea6a909.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d341e209131cea38614.md>)

Recorded evaluation configuration: {"Backbone":"ViT","Optimization Epochs":"25"}. This record does not assert independent reproduction.

### Naive

AA: 17.56. Result ID: 69ef008847f34ce924d470a4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"8","Privacy parameter (delta)":"1e-5","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### Naive

AA: 15.93. Result ID: 69ef008847f34ce924d470a3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"1","Privacy parameter (delta)":"1e-5","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

### Naive

AA: 13.15. Result ID: 69ef008847f34ce924d470a5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78bb1e209131cea7a1ca.md>)

Recorded evaluation configuration: {"Differential Privacy budget (epsilon)":"non-DP","Backbone":"ViT-B","Pre-training dataset":"ImageNet-21k"}. This record does not assert independent reproduction.

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