# Image Classification on CIFAR-100 20

Source: [Image Classification on CIFAR-100 20](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20>)

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

Data updated: 2026-05-21T11:48:19.013Z

Evaluation of Image Classification performance on the CIFAR-100 20 variant, reporting accuracy across five specific sub-tasks (Dtar,1 through Dtar,5) and the overall average accuracy.

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) | CIFAR-100 20 |
| Metric | Dtar,1 Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [Dtar,1 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20.md?metric=Dtar%2C1%20Accuracy>): Higher is better; unit: PERCENT.

- [Dtar,2 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20.md?metric=Dtar%2C2%20Accuracy>): Higher is better; unit: PERCENT.

- [Dtar,3 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20.md?metric=Dtar%2C3%20Accuracy>): Higher is better; unit: PERCENT.

- [Dtar,4 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20.md?metric=Dtar%2C4%20Accuracy>): Higher is better; unit: PERCENT.

- [Dtar,5 Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20.md?metric=Dtar%2C5%20Accuracy>): Higher is better; unit: PERCENT.

- [Average Accuracy](<https://www.sota2.com/research/sota/image-classification-on-cifar-100-20.md?metric=Average%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Dtar,1 Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Tunable MAGMAX (Label) | 81 | Not recorded | 6a0e9173f984739ed67389bc |
| Tunable MAGMAX (OT) | 79 | Not recorded | 6a0e9173f984739ed67389bc |
| Baseline | 78 | Not recorded | 6a0e9173f984739ed67389bc |
| MAGMAX | 78 | Not recorded | 6a0e9173f984739ed67389bc |
| Random Mix | 77 | Not recorded | 6a0e9173f984739ed67389bc |
| Average (Model Soup) | 77 | Not recorded | 6a0e9173f984739ed67389bc |
| TIES-Merging | 77 | Not recorded | 6a0e9173f984739ed67389bc |

### Tunable MAGMAX (Label)

Dtar,1 Accuracy: 81. Result ID: 6a0ef0fcfbc373592c0747bf.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

Recorded evaluation configuration: {"Preference Control":"label distribution-based preference control (Label)","Scaling factor (lambda\_merge)":"0.5"}. This record does not assert independent reproduction.

### Tunable MAGMAX (OT)

Dtar,1 Accuracy: 79. Result ID: 6a0ef0fcfbc373592c0747be.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

Recorded evaluation configuration: {"Preference Control":"OT distance-based preference control (OT)","Scaling factor (lambda\_merge)":"0.5"}. This record does not assert independent reproduction.

### Baseline

Dtar,1 Accuracy: 78. Result ID: 6a0ef0fcfbc373592c0747b9.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

Recorded evaluation configuration: {"Description":"Sequentially fine-tunes the model on each task without any mitigation of CF"}. This record does not assert independent reproduction.

### MAGMAX

Dtar,1 Accuracy: 78. Result ID: 6a0ef0fcfbc373592c0747bd.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74521e209131cea610c2.md>)

Recorded evaluation configuration: {"Scaling factor (lambda\_merge)":"0.5"}. This record does not assert independent reproduction.

### Random Mix

Dtar,1 Accuracy: 77. Result ID: 6a0ef0fcfbc373592c0747ba.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74521e209131cea610c2.md>)

Recorded evaluation configuration: {"Scaling factor (lambda\_merge)":"0.5"}. This record does not assert independent reproduction.

### Average (Model Soup)

Dtar,1 Accuracy: 77. Result ID: 6a0ef0fcfbc373592c0747bb.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ce11e209131cea3679c.md>)

[Implementation](<https://github.com/mlfoundations/model-soups>)

Recorded evaluation configuration: {"Scaling factor (lambda\_merge)":"0.5"}. This record does not assert independent reproduction.

### TIES-Merging

Dtar,1 Accuracy: 77. Result ID: 6a0ef0fcfbc373592c0747bc.

[Reporting paper](<https://www.sota2.com/research/paper/6a0e9173f984739ed67389bc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71091e209131cea4ecd8.md>)

Recorded evaluation configuration: {"Scaling factor (lambda\_merge)":"0.5"}. This record does not assert independent reproduction.

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